Published July 1, 1969
The Transport of Carbohydrates
by a Bacterial Phosphotransferase System
SAUL ROSEMAN
From the McCollum-PrattInstitute and the Department of Biology,Johns Hopkins
University, Baltimore, Maryland 21218
This paper is concerned with a bacterial phosphotransferase system; its
properties, and evidence indicating it to be responsible for sugar transport in
bacterial cells will be briefly reviewed, and some speculations will be offered
concerning the physiological implications of sugar transport via this system.
T h e discovery of the phosphotransferase system resulted from our longstanding interest in the biosynthesis of carbohydrate containing macromolecules (1). T h e 9-carbon sugar acid, sialic acid, is a frequent component
of these macromolecules, and enzymatic degradation of one of the sialic acids
(N-acetylneuraminic acid) was found to give pyruvate and N-acetyl-Dmannosamine (2). Studies on the metabolism of the latter sugar led to the
discovery of a specific kinase that catalyzes the reaction shown in Fig. 1 ; this
kinase is widely distributed in animal tissues (3). Since certain bacterial cells
synthesize polymers of N-acetylneuraminic acid, or can metabolize N-acetylD-mannosamine, extracts of these cells were examined for the kinase. T h e
sugar was not phosphorylated by the reaction shown in Fig. 1, but it was
phosphorylated when phosphoenolpyruvate (PEP) was substituted for ATP.
T h e bacterial system, designated PEP:glycose phosphotransferase system, or
simply phosphotransferase system, was found to catalyze the reaction shown
in Fig. 2 (4, 5).
PROPERTIES
OF
THE
ENZYME
SYSTEM
Components of the Phosphotransferase System
T h e complete phosphotransferase system (designated in some of the Figures
as I T S ) catalyzes the reaction shown in Fig. 2. Phosphoenolpyruvate is the
~38 s
The Journal of General Physiology
Downloaded from on June 18, 2017
ABSTRACT The components and properties of a phosphoenolpyruvate: glucose phosphotransferase system are reviewed, along with the evidence
implicating this system in sugar transport across bacterial membranes. Some
possible physiological implications of sugar transport mediated by the phosphotransferase system are also considered.
Published July 1, 1969
SAUL ROSEI~N Bacterial PhosphotransferaseSystem
i39
s
phosphate donor, and a variety of sugars can serve as acceptors. While most
of the enzymatically synthesized monosaccharide phosphate esters have been
shown to be the 6-phosphate esters (4, 5), including the products from
galactose (6), and the lactose analogue thiomethyl /3-D-galactopyranose (7,
8), recently, the product obtained from fructose has been identified as fructose1-P (9, I0). The products formed from disaccharides have not yet been characterized with the exception of lactose-P, which is synthesized by extracts
obtained from Staphylococcus aureus, and here the phosphoryl group is linked
CHO
CHO
+
ATP
+ ADP
~,
CHz-O-PO~
N-ACYL-D- MANNOSAM/NE
N-ACYL-D-MANNOS,4MINE
6- PHOSPHATE
R = ACETYL OR GLYCOLYL
Inactive:
N-AcyI-Gm, N-Ac-GoIm, Mm, Gm, Golm, Mannose, Glucose, Galoctose,
Fructose, GTP, UTP, CTP, TTP, ITP, dATP, PEP
FiotmE 1. N-acyl-D-mannos~mlne kinasc (rat liver: 2000-fold). (Ghosh, S., and S.
Roseman. 1961.Proc. Nat. Acad. Sd. U.S.A.; Warren, L., and H. Felsenfeld. 1961.Biochcm.
Biophys. Res. Commun.; Kundig, W., S. Ghosh, and S. Roseman. 1964.J. Biol. Chern.)
Sugar
+
Mg++
CH z = C - C02
I
_
O-PO~"
PTS
~
Phosphoenolpyruvote
Sugar - PO;
+
CH3COC0~_
Pyruvo/e
(PEP)
FIGURE 2.
Bacterial phosphotransferase reaction.
to C-6 of the galactose moiety (11). T h e phosphotransferase system requires
PEP as the phosphate donor (Fig. 3), which makes the system different from
uther known sugar kinases. Further evidence that PEP itself is the donor, and
no~ a nucleotide triphosphate generated from PEP, has been published (4).
"1he complexity of the phosphotransferase system is illustrated in Fig. 4.
In the case of Gram-negative organisms (Escherichia call, Salmonella typhimurium), the system consists of three protein fractions, Enzyme I, Enzyme II,
and a low molecular weight histidine-containing protein, designated HPr.
In the Gram-positive organism Staphylococcus aureus, an additional protein
Downloaded from on June 18, 2017
CH20H
Published July 1, 1969
i4o
s
TRANSPORT
PROTEINI
component has recently been isolated from cells induced to grow on lactose
or mannitol, and the new protein is called Factor III (12). Factor III obtained from mannitol-induced cells, is not the same as the Factor III from
lactose-induced cells, each Factor III being required for phosphorylation of
the sugar used for growth of the cells from which it was isolated. A protein
component distinct from Enzymes I, II, and HPr, has recently been detected
in extracts obtained from Aerobacter aerogenes (9, 10), and is required for
Active
PEP
K m 4x
I0 -4 M
Inactive
P - Histidine
and Triphosphates
Cyclic 3 ' - 5 * A M P
P - Serine
N - Phospho - Glycine
Creatine - P
Phosphoramidate
ATP, A D P _+ Pyruvate Kinase
P - Glycerate
P P i , Pi
C o A + Succinate + Pi
Acetylphosphate
TPN, TPNH
I
Fraction
Enzyme -
Glucose - 6 - P
Substrate specificity: P-donor.
FmuR~. 3.
Protein
DPN, DPNH
I
Occurrence
Sugar
Specific
Inducibility
Cytoplasm
-
-
HPR
Histidine - Protein
Cytoplasm
-
-
Tr
Fraction or Enzyme I I
Membrane
+
+ and -
Cytoplasm
+
+
III
Factor
]'II (,5". oureus)
FIOURE 4.
Components of bacterial PTS.
phosphorylating fructose when the concentration of this ketose is low. The
authors therefore consider that the protein functions by increasing the affinity
of Enzyme II for fructose.
The requirements for the different protein fractions in the phosphorylating
system are illustrated in Figs. 5 and 6. The cellular location of the various
proteins is also given in Fig. 4; these conclusions are based on the results of
lysis experiments, followed by subcellular fractionation (5). All of the detectable Enzyme II is found in the membrane fraction, while more than 95 %
of the other proteins are found in the supernatant fluid. Enzyme II appears to
Downloaded from on June 18, 2017
Nucleoside Mona -, Di-,
Published July 1, 1969
SAUL ROSE,S_AN Bacterial Phosphotransferase System
x41 s
be an integral component of the membrane, since exhaustive washing of the
membrane, or its rupture with either the French Press or by ultrasonic vibration, does not solubilize this component. On electron microscopic examination, the Enzyme II preparations appear to be very similar to the membrane
vesicles studied by Kaback and Stadtman (13).
PEP + N-Acetylmonnosomine " - ~ ' N - A c M m - 6 - P + Pyruvote
N- Ac-Mm- 6- P
for reed~
/zmo le
incubation
0.38
Minus Mg ++
0.03
Minus
I
0.00
Minus
TI"
0.00
Minus
HPR
0.00
~ 3 0 rain , 37°C :, 20p.g HPR',5pg each, I ondII
F m u ~ 5.
Requirements for Esclwrichia call PTS.
PEP + T M G ' - - 4 " T M G - 6 - P + Pyruvote
Incubation
14C-TMG-P formed
cpm
"1" II, ]II, HPR, Mg++
Minus
Minus
Minus
Minus
Minus
Mg ++
I
II
HPR
m
15,700
150
B.00
300
460
900
Lactose operon induced by growth on galoctose
TMG = methyl / ~ FIo~
6.
thiogoloctoside
Requirements for Staphylococcus aureus PTS.
The remaining points in Fig. 4 concern sugar specificity, and inducibility
of the protein components of the phosphotransferase system. As will be
emphasized later, the specific sugar requirements of the system are confined
to Enzyme II (and to Factor I I I where it is required), while Enzyme I and
H P r are common to all sugars phosphorylated by the system. The inducibility
of some components of the system is shown in Fig. 7 (12). When S. aureus is
grown under inducing conditions for the lactose operon, and extracts of these
Downloaded from on June 18, 2017
I,II, HPR, Mg ++
Published July 1, 1969
TRANSPORT PROTEINS
I42 s
cells are compared with extracts of noninduced cells, there is no detectable
increase in Enzyme I and H P r upon induction, while there are dramatic
increases in Enzyme II and Factor III. (In these experiments, the lactose
analogue thiomethyl ~-D-galactopyranoside was used as the substrate; this
compound, designated T M G , is often used in both transport and phosphorylation experiments since it is not metabolized, and it is not cleaved by"
/~-galactosidases.) T h e results therefore show that Enzyme I and H P r are constitutive, while Factor III, and at least one Enzyme II are inducible. Experiments with cells grown on different sugars indicate that most Enzymes II are
inducible, except for those which phosphorylate D-glucose and its analogues
(such as methyl/3-D-glucopyranoside), and D-mannose and its analogues (such
as N-acetyl-D-mannosamine). I n some cases, when cells are grown on a
glucose mineral salts medium, the extracts will also phosphorylate galactose
Specific activity in Extracts
Noninduced
Z
HPR
52
Induced
58
6.8
II
0.1
Ill
O. I
6.7
38
2.5
Spec. act.= p.moles TMG-P formed ( x 1 0 2 ) / m g protein in
crude extracts; 30 rain at 37°C. Cells grown in boctopeptone+_.
galactose or G o I - 6 - P .
FIOU~ 7.
Induction of Enzyme I I and Factor I I I (S.
aureus
PTS).
and fructose at low rates, but these rates are substantially increased when the
cells are grown on the respective sugars, or the enzymes are derepressed by
growth in the absence of glucose.
Attempts are in progress to completely resolve the proteins from each other,
and to obtain each in homogeneous form. H P r has been obtained in homogeneous state, and is discussed below. E n z y m e I from E. coli has been purified
about 500-fold (Saier et al., unpublished results). Factor I I I from S. aureus
(lactose-induced cells) has been partially purified. The complex called Enzyme
II is discussed below.
Partial Reactions of the Phosphotransferase System
Using the purified proteins, it has been possible to demonstrate that the
reaction shown in Fig. 2 is the net reaction of the two steps shown in Fig. 8
(4, 12). I n the first step, Enzyme I catalyzes the transfer of phosphate from
PEP to HPr, giving pyruvate and phospho-HPr. In the second step, E n z y m e
Downloaded from on June 18, 2017
Component
Published July 1, 1969
i43 s
SAUL ROSEMAN BacterialPhosphotransferase System
II (in the presence of Factor III) (12) catalyzes the transfer of phosphate from
phospho-HPr to the sugar, regenerating HPr. The histidine-containing protein therefore serves as a phosphate carrier.
Some of the evidence supporting the formulation of these reactions is
shown in Figs. 9 and 10.
PEP
+ HPR
Phospho-HPR
I
-..,--Mg++--
Phospho-HPR + Pyruvote
+ Sugar
H
I I I --- S u g a r - P
l,Mg ++
PEP + Sugar
,-
+
HPR
S u g o r - P + Pyruvote
I I , ]]1", H PR
Factor ]]3" detected in induced
S. oureus
3z p EP
Enzyme I
+
---------,--
Histidine - Protein
Enzyme
32p _ Histidine - Protein
Mg ÷+
Pyruvate
3ap Incorporated into Protein
cpm
I + H -P
129,000
I + IT + HP
118,000
:E + HP
200
I , 11" or HP
150-200
Incubations (0.S ml) : H - P O.Omg, 3RPEP 2.5 /zm (sp.A. IIALc /,mm)
MgClg 2.S/zm, Tris-HCI pH 7.4 25.0,ALto,
Enzymes 0.5 rag; 30min, 37°C
FIOURE 9.
Phosphorylation of HPr by Enzyme I.
Properties of HPr and Phospho-HPr
HPr has recently been obtained in homogeneous form from E. coli (7). T h e
purification procedure includes a variety of standard methods for protein
purification; the initial step involves heat treatment (100°C, 10 rain) of the
crude extract, while the last step involves chromatography on DEAE-cellulose.
T h e fractionation on DEAE-cellulose gives two symmetrical protein peaks,
exhibiting H P r activity; HPr-1 is eluted from the column before, and is more
active than HPr-2. On examination of the two proteins by polyacrylamide
disc gel electrophoresis at four p H values, both proteins appear to be homogeneous, but HPr-2 migrates more rapidly than HPr-1 at p H 9.5. Further
Downloaded from on June 18, 2017
FIGURE 8. Reactions of the phosphotransferase system.
Published July 1, 1969
I44 s
TRANSPORT
PROTEINS
examination of the proteins shows that they contain the same amino acids in
the same proportions, do not contain detectable quantities of carbohydrate
or phosphorus, and are homogeneous and exhibit the same molecular weight
in the ultracentrifuge. The only detectable difference between the two proteins is in their amide content; HPr-1 contains 15 amide residues per mole,
while HPr-2 contains 10-11. This difference in amide content explains the
difference in electrophoretic behavior, and the order of elution from the
DEAE-cellulose column. Further investigation suggests that HPr-1 is converted to HPr-2 during the first step in the isolation procedure; i.e., the heat
step results in the loss of some amide residues from HPr-1. Thus HPr-2
appears to be an artifact of the isolation procedure. We now believe that
HPr-1 is not the native histidine protein. Unheated crude extracts yield a
32 p _ H i s t i d i n e - P r o t e i n
+
E n z y m e 1T
,=
Mg ++
Enzyme
Histidine -
N-Ac-Mrn
- 6 - 32 p
Protein
- 6-32p
cpm
lz
17,4oo
I + H
17,000
I
40
--
30
Incubotions (0.S ml) .. 3e P-H-P 22,500 cpm , N - Ac-Mm 4.0 ,xJm,
MgC/2 2.5 /Jm, Tr/s-HCI pH 7.4 25.0/Jm,
Enzymes 0.5 mg; 30 min, 37°C
FIGURE 10.
82p transfer from 82P-H-P to N-acetylmannosamine.
protein with HPr activity that is eluted from the DEAE-cellulose column
earlier than HPr-1. Attempts to isolate the native material are in progress.
The properties of HPr are summarized in Fig. 11. The amino acid composition is unusual in that cysteine, tyrosine, and tryptophane are absent. In
fact, the purified protein exhibits an ultraviolet absorption spectrum similar
to that obtained with phenylalanine (it contains 4 moles phenylalanine).
HPr also contains two histidine residues. Stoichiometry studies with both
HPr-1 and HPr-2 show that 0.82-0.95 mole of ~P can be transferred to 1
mole of HPr, using prolonged incubation times, and excessive quantities of
8*P-labeled PEP and Enzyme I. Our earlier studies (4) had demonstrated
that the phosphoryl residue was linked to histidine, and the stoichiometry
experiments therefore suggest the possibility that only one of the two histidine
residues accepts phosphate.
Downloaded from on June 18, 2017
N - Acetylmannosamine
N - Acetylmonnosamine
+
Published July 1, 1969
SAUL RO$EMAN BacterialPhosphotransferase~ystem
145
s
Some properties of phospho-HPr are summarized in Fig. 12. T h e phosphoryl group is linked to N-1 of the histidine imidazole ring (7). This linkage
is unlike that found in succinate thiokinase [the phosphohistidine-containing
protein first observed by Boyer et al. (14), and D e L n c a (15)] in which the
phosphate is bound to N-3 of the imidazole ring of a histidine residue. Identification of the attachment site of the phosphoryl group was m a d e possible by
the synthesis of the N-1 and N-3 isomers of phosphohistidine by Hultquist
I.
Homogeneous
o. Disc gel electrolohoresis;pH 2.3, 4 . 3 , 6 . 6 , 9 . 5
b . Ultrocentrifuge
2.
Moleculor
9340
Weight
by ultrocentrifugolion (Yphontis}
+_ 2 3 0
3.
by T.-omino
ocid residues
Anolysis
o . No corbohydrote or P.
b . Usuol
omino ocids except Cys, T y r , Trp
c . 2 His/mole
FIOURE l 1. Someproperties of E. coli HPr.
I
PEP
+
HPR
...-Mg++ Phospho-HPR
I mole P incorporoled//mole
P linked
to N - I
~A~,NH
N~~,,,.N- CHz
CI H - PCO
O~"VVk,
I - P - Hislidine
of
+ Pyruvote
HPR
His imidozole ring
~XA'
~0P -NN,,~N
H~- CHCHz
I C0~/V~l
3 - P - Histidine
Floux~ 12. Phospho-HPr (E. coli).
et al. (16, 17). T h e phosphoryl-N bond in these substances is very labile to
acid, and the N-1 isomer is considerably more sensitive than the N-3 isomer.
The Enzyme H Complex
As indicated earlier, m a n y Enzymes II are inducible, while several are found
in cells grown on glucose, and are considered constitutive. How m a n y different
Downloaded from on June 18, 2017
9639
Published July 1, 1969
TRANSPORT PROTEINS
~46 s
Enzymes II can be induced in a single organism? We do not know, but believe
that there are many such enzymes, specific for different sugars. For example,
the Enzyme II (membrane) fraction from cells grown in a rich m e d i u m (broth
containing high concentrations of glucose) catalyzes the phosphorylation of
glucose and mannose and their analogues, but with little or no activity on a
large number of other sugars (4). The sugars that are inactive with this
type of preparation are, however, active acceptors when the cells are grown
in a mineral salts m e d i u m containing the sugars per se. Compounds that can
be phosphorylated include pentoses, pentitols, hexoses, hexitols, disaceharides,
etc.
Incubation
Mixture
/zmoles 6 -Phosphate formed (x 102)
N-AcMm
10.8
12.5
Minus l"r_.-A
0.0
0.0
Minus ~ - B
0.4
2.1
1.4
0.4
II-A, II-B
.I
]'r'-Lipids,[ P-HPR] ;
Minus 17- Lipids
*LP-HPRJ"
" generating system=Enzyme I , HPR, PEP, Mg ++ ,
phosphate pH 7.4.
Incub. 60min ,37°C,Proteins~-A and I I - B ,
0.1 rng each ~ E co// crude lipids or 40 Fg phosphotidyl-glycerol.
All values corrected for (Minus PEP)control, opproxirnotely 1.0 .
F i a u ~ 13. Required components of E. coli constitutive Fraction II.
The Enzymes II from E. coli grown on glucose have recently been solubilized (18). Fractionation of these extracts show that three components are
required for phosphorylating activity, two protein fractions (II-A and II-B),
and a lipid fraction (II-lipid). Omission of any of these fractions from incubation mixtures containing a phospho-HPr-generating system gives little or no
sugar phosphate (Fig. 13). Further study suggests that II-A contains two
specific proteins, one active for glucose and the other for mannose phosphorylation.
The lipid requirements for the reconstituted system have recently been
defined. Substitution of a variety of lipids for II-lipid shows that small quantities of phosphatidyl-glycerol (kindly provided by Dr. William Lennarz) are
essential for activity.
Purification of II-B has been effected to the extent that it yields a single
protein band in sodium dodecyl sulfate, urea polyacrylamide disc gel electrophoresis, and the molecular weight of this protein is approximately 35,000
Downloaded from on June 18, 2017
a- MeGlu
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
I47 s
(19). Homogeneity is not claimed on this basis, since a mixture of proteins
containing subunits of about the same size would yield the same results.
Finally, an observation that is potentially of great importance concerns the
order of mixing II-A, II-B, II-lipid, and divalent metal. O p t i m u m activity is
only observed when a particular sequence of mixing is followed. These results
suggest the possibility of studying the requirements for reconstituting intact
and functional membrane, and for determining how the proteins interact
with m e m b r a n e lipids and metal ions.
ROLE
OF PHOSPHOTRANSFERASE
IN TRANSPORT
SYSTEM
It is our contention that the PEP:glycose phosphotransferase system is
intimately involved in the transport of sugars across bacterial cell membranes.
I ,±
Active
--S
Tronsport
Energy
'--S-X
Group Tronsloco/ion
Fmue.E ]4.
Transport systems.
In this section we will present a simple model that suggests how the system
m a y function, and then review some of the evidence in its support.
Several types of transport systems have been defined, and their relationships have been reviewed (20-23). Only the following types will be considered
in this discussion:
Simple Diffusion
This process involves the diffusion of solutes across
membranes, where neither the m e m b r a n e nor any component thereof acts
in a catalytic m a n n e r as a "carrier" to expedite the process. W h e n simple
diffusion occurs in living cells, it resembles the transport of such solutes across
synthetic hydrophobic membranes. Simple diffusion exhibits the following
characteristics: (a) T h e process depends upon the existence of an electrochemical gradient across the membrane, and it tends to make the gradient
disappear. (b) It is not a stereospecific process. For example, L-glucose and Dglucose would penetrate the m e m b r a n e at the same rate. (c) T h e rate of
Downloaded from on June 18, 2017
I
Focilitoted
Diffusion
Published July 1, 1969
x48 s
TRANSPORT
PROTEINS
penetration is a function of the gradient and of other parameters such as the
thickness of the membrane, diffusion coefficient, etc. (d) W h e n all other factors
are constant, the rate of penetration is directly related to the concentration
gradient; i.e., the rate does not approach a limiting value, Vm~xwith increasing
concentration of substrate. Other features of simple diffusion, such as solvent
flow or " d r a g , " have been reviewed (22). Simple diffusion of sugars across
bacterial membranes is not thought to be a significant physiological process,
either in the entry or exit mechanisms.
Active Transport is similar to facilitated diffusion, except that the solute is
moved across the m e m b r a n e against the concentration gradient. This process
requires the expenditure of metabolic energy.
Group Translocation will be used here to mean the process shown in Fig.
14, in which penetration of the m e m b r a n e by a solute occurs concomitant
with a chemical reaction resulting in the formation of a derivative, illustrated
in the figure as S-X. Metabolic energy is also required for group translocation,
and we believe that this is the most general mechanism for the transport of
sugars across bacterial membranes.
T h e transport systems enumerated above encompass the different bacterial
permease systems. There is, thus far, no evidence for coupled transport systems
of the type found in animals (22), where sugars and amino acids are cotransported with sodium ion, and the energy for transport m a y be derived from the
sodium pump.
Downloaded from on June 18, 2017
Facilitated Diffusion and two other processes of interest in this discussion
are illustrated in Fig. 14. As indicated in the figure, facilitated diffusion is
similar to simple or passive diffusion in the sense that the solute moves down
a concentration gradient, and the process does not require the expenditure of
metabolic energy (except for the energy required to maintain an intact
membrane). It differs from simple diffusion as follows: (a) T h e m e m b r a n e
contains a component (P in Fig. 14) which catalyzes the process, and the rate
of the facilitated diffusion process is m u c h more rapid than predicted for
simple diffusion. (b) T h e catalytic component, P, is generaUy stereospecific.
This statement implies the presence of different catalytic systems, or perincases, for different solutes, and this fact was demonstrated in bacterial cells
m a n y years ago by Rickenberg, Cohen, Buttin, Monod, Kepes, and their
collaborators at the Pasteur Institute (23). (c) T h e rate of penetration generally approaches a limiting value with increasing concentration of the solute
on one side of the membrane. This value is called V ~ . T h e rate of entry as
a function of solute concentration resembles simple Michaelis-Menten enzyme
kinetics, and transport data are frequently handled in the same manner; i.e.,
by using Lineweaver-Burke plots to determine the Kr, of the permease system
for the solute, and to determine the V~m~
y.
Published July 1, 1969
SAUL ROSEMAN
Bacterial Phosphotransferase System
z49 s
In general, wild-type bacterial cells contain a constitutive glucose permease
system, but the transport of the few other sugars that have been studied
appears to proceed via inducible systems (23). Where possible, transport is
measured with nonmetabolizable analogues of the sugar under study, so that
accumulation of metabolites by the cell is not confused with accumulation of
the solute under study. For this reason, methyl ~-D-glucopyranoside is generally used as a glucose analogue, methyl/3-D-galactopyranoside as a galactose
analogue, and methyl /3-thiogalactopyranoside (TMG) as an analogue of
lactose (23).
GROUP TRANSLOCATION
Outside
Membrane
Inside
riG'" / I
-I1 P-H
. PR,I HPR
s.
E. coli
Stop& oureu$
I
S. #yphimurium] mutants (HPR, I )
E. coli
J
IG-~ -,,-- ~ ] I G o I
:r q
ACTIVE TRANSPORT
S. typhimurium
E. col~
{'Jar'°°"/
Fmum~ 15. Sugar transport via the phosphotransferas¢ system.
Relationship between the Phosphotransferase and Permease Systems
A simple model to explain how the phosphotransferase system can function
in the transport processes in Fig. 14, is given in Fig. 15. We believe that most
sugars penetrate bacterial membranes by group translocation, mediated by
the respective Enzyme II for each sugar. T h e sugar is phosphorylated as it is
transported, and the phosphate is derived from phospho-HPr.
Another mechanism, facilitated diffusion, does occur in certain mutants
described below, and with a few sugars. Finally, the active transport process
also takes place in a few systems, and again it is proposed that the process is
mediated by the specific Enzymes II, and the energy derived from phosphoH P r (i.e. PEP). T h e sugar may be phosphorylated in the membrane, and the
phosphate ester then hydrolyzed by sugar-specific membrane-bound phosphatases of a type that have been described (24), or the Enzyme II-sugar
complex changed to a phosphorylated complex (II' in the figure), which has
a low affinity for the sugar, and releases it to the interior of the cell. T h e latter
mechanism would fit the p u m p mechanism proposed by Kepes (23).
Downloaded from on June 18, 2017
FACILITATED DIFFUSION (few sugors)
Published July 1, 1969
TRANSPORT PROTEINS
15o s
T h e model offered in Fig. 15 is purely speculative. What, in fact, is the
evidence that the phosphotransferase and permease systems are at all related?
There are six separate lines of evidence, and these are summarized in Fig. 16.
The major emphasis in this brief review will be on the most important correlation, the genetic studies.
While detailed kinetic studies are yet to be performed, there is a good correlation in the action of inhibitors on specific permease and phosphotransferase
systems; the inhibitors tested include metal ions, sugars, and sulfhydryl
reagents. For example, phosphorylation of galactose and T M G by appropriately induced E. call membrane preparations is inhibited by glucose, while
galactose is a particularly potent inhibitor of T M G phosphorylation; these
results correspond to the inhibition of transport of the respective sugars in
intact E. cdi cells (25).
HPR restores transport in osmotically-shocked E. co/i.
2.
Sugar-P are accumulated during transport in whole cells.
3.
Genetic correlations
4.
Identity between inhibitors of PTS end permeose systems.
5.
Specificities of Fractions 1I correspond to permeases.
6.
Sugar uptake by membrane vesicles (Kabock)=
(a}PEP dependent (b) accumulates Sugar-P.
FXGURE 16.
Correlations between PTS and sugar permease systems.
As indicated earlier, most Enzymes II are inducible, but some, such as the
Enzyme II that phosphorylates glucose, are constitutive. T h e same pattern
holds for the sugar permease systems. In addition to this qualitative correlation, induction of specific permeases is accompanied by a concomitant quantitative increase in the activity of the corresponding Enzyme II (and Factor
I I I ; Fig. 7).
Finally, Kaback has utilized membrane vesicles from E. call to study sugar
uptake (26). He finds that the membranes are capable of taking up sugars,
provided that they are supplemented with PEP; the sugars are accumulated
in the vesicles as the phosphate esters. Furthermore, using double-labeling
techniques, the sugar-phosphate ester accumulated by the vesicles was shown
to be formed during transport, rather than after penetration into a pool of
the free sugar inside the membranes.
Restoration of Transport in Osmotically Shocked Cells by HPr
In these experiments (27), we discovered that the osmotic shock procedure of
Nossal and Heppel (28), resulted in a marked decrease in the ability of the
Downloaded from on June 18, 2017
I.
Published July 1, 1969
SAULROSEMAN BacterialPhosphotransferaseSystem
x5x s
cells to transport sugars, despite the fact that the cells remained viable.
Futher study showed that the cells also lost most of their HPr. When the
shocked cells were first incubated with purified HPr, the ability to transport
T M G and methyl a-glucoside was restored. Fig. 17 shows the results obtained
with T M G in E. coli (i+ y+ z-) induced for the lactose operon. A variety of
experiments showed that the effect was very specific for HPr, and the cell
membranes had not lost their specificity for TMG.
The model in Fig. 15 shows HPr on the inside of the membrane (or perhaps
in the membrane). Incubation of the shocked cells with relatively high concentrations of HPr would therefore require (to fit the model) that a small
quantity of HPr either penetrates the membrane, or is absorbed to it (by
forming a complex with Enzyme II?) in order to be functional.
I
o
x
I
l
I
I
i
I
I
I
I
14
12
E
fx
o.
o
I0
_o
8
X
.--.-x
~
' 'X'
'X
Shocked + HPR
Norrnol
'
o
o
4
~O'o~O~O
I-i
<.}
0
0
*0
Shocked
HCHO- Control
0
I
4
I
8
I
12
l
16
I
20
I
24
I
28
I
32
I
36
I
40
I
44
TIME (MINUTES)
FiOtrl~ 17.
Restoration of TMG transport in osmoticallyshockedE. coliby HPr.
Accumulation of Sugar Phosphates during Transport
To our knowledge the first demonstration that sugars were accumulated as
their phosphate esters during transport was that of Rogers and Yu (29); the
sugars were galactose, which accumulated as the 6-phosphate ester, and aand fl-methylglucosides. Subsequent work by other laboratories showed that
essentially all of the methyl a-glucoside taken up during the 1st min by E. coli
was detected as the phosphate ester. One of these experiments (30) is shown
in Fig. 18. In S. aureus, it appears that essentially all sugars are converted to
the phosphate esters during transport (8, 11, 31).
Downloaded from on June 18, 2017
?
I
Published July 1, 1969
z52 s
TRANSPORT
PROTEINS
Genetic Correlations
The most important evidence relating the phosphotransferase and permease
systems has been obtained with mutants. As will be seen, the observed physiological behavior of the mutants conforms to the predicted behavior implicating the phosphotransferase system in the transport process.
The model relating the phosphotransferase and transport systems in Fig.
15, has been slightly modified in Fig. 19, primarily to emphasize again that
the specificity of each sugar permease system is derived from the specificity
of the different Enzymes II. Fig. 19 also shows that most sugars are thought
to be transported by group translocation, while some may be transported by
Total
14C - MeGlu
!
15
E
5t- ~o
O
b
2
Fmug~ 18. a-Me-glucosideuptake by E. coli (Winkler,
,b
I
2o
,'o
!
5o
1966).
minutes
active transport. Finally, the figure shows another critical feature of this
model, namely that Enzyme I and HPr are required for all sugars transported
via the phosphotransferase system when it catalyzes either group translocation
or active transport (facilitated diffusion will be considered below).
The model in Fig. 19 leads to certain obvious predictions, which are illustrated in Fig. 20. (a) In mutants containing a defective Enzyme II specific
for a particular sugar, then only the transport of that sugar should be affected.
(b) In mutants defective in either Enzyme I or HPr, the group translocation
and active transport of all sugars transported by the phosphotransferase
system should be affected. As shown in the figure, Enzymes II in mutants
defective in Enzyme I or H P r should be present in the membranes, as they
are in the wild type. Thus, it is possible that facilitated diffusion could occur
Downloaded from on June 18, 2017
-Phosphot[
Published July 1, 1969
SAUL ROSEMAN Bacterial PhosphotransferaseSystem
x53 S
in E n z y m e I or H P r mutants. Whether or not facilitated diffusion will occur
depends upon the exact mechanism required for the transport of each sugar.
In the simplest model, we might predict two possibilities, one that the dissociation between the E n z y m e I I - s u g a r complex on the inside of the m e m b r a n e
proceeds at a reasonable rate and facilitated diffusion therefore takes place
(St in Fig. 20), or, it does not (S~. in Fig. 20).
Out
Membrane
In
S,
P-HPA ~
Group
Tronslocotion
~iz
S2
Sz--P
I
I
]In
Sn
"
Sn-p
llx~ sxt
P-HPR-.~--
HPR + PEP
/
Fzoum~ 19. Sugar transport in wild-typecells.
Enzyme ]I
Ou! I
I In
Enzyme I or HPR
I. All sugars (S t ........ Sn)
2. Facilitateddiffuslon
occur
S I~ ] , Ill l].~ B i
S.+IV]]
FiGtr~ 20. Predicted transport defects in PTS mutants.
T h e properties of mutants defective in specific Enzymes II, listed in Fig.
21, agree with their predicted properties. In addition to the reported results,
detailed kinetic studies have recently been performed in this laboratory
(Simoni et al., unpublished results) with E n z y m e II and Factor I I I mutants
of S. aureus (lactose operon). T h e transport studies showed that such mutants
were unable to take up T M G regardless of the external concentration, and
therefore that neither the process of passive nor facilitated diffusion occurs at
Downloaded from on June 18, 2017
Sx
HPR + PEP
II
II
Active
Transport
I
Published July 1, 1969
TRANSPORT
154 s
PROTEINS
significant rates in such cells. T h e mutants did take up methyl a-glucoside at
rates comparable with those of the parent strain.
T h e second predicition in Fig. 20 concerns the properties of Enzyme I and
H P r mutants. As shown in Fig. 22, several such mutants have been reported.
T h e original S. aureus m u t a n t of Egan and Morse (31), which was unable to
utilize 11 carbohydrates for growth and was shown to be defective in its
ability to transport the sugars, has recently been shown to be an Enzyme I
Carbohydrate
Affected
Organism
Growth
Transport
A. oerogenes
Mannitol-
Monnitol-
E. co//
Lin(1968)
/~-Glucosides- ,8-Glucosides- F'ox,Wilson(1968)
Factor ~ I . -
Loc-
Loc-
Morse et al. (1968)
mutants behave similarly
FIOLrRE 2 l. Specificcarbohydrate defects in Enzyme II mutants.
Organism
Carbohydrates Affected~
Growth Transport
E. co//
I-
5
r
Tanaka et al. (1967)
A. aerogenes
I - or HP~
5
y
Tanaka, Lin (1967)
S. typhimurium
1'- or HPR"
9
negative
Simoni et al. (1967)
S. aureus
I-
I1
negative
Egan, Morse (1965)
E. co#
I-
12
negative
Fox, Wilson (1968)
Includes glycerol, hexoses (GIc, Gal, Man), glucosides, GIcNAc,
ketohexoses, pentoses (t,) , disaccharides (Lac, Mal, Suc, Mel)
Fzotrm~ 22. Multiple sugar defects in Enzyme I and HPr mutants.
m u t a n t (12). This, and the other mutants listed in the figure, are unable to
grow on, or ferment, from 5 to 12 carbohydrates. T h e defect was shown to be
in transport in the last three cases shown in the figure, b u t was not studied in
the first two (31-35). T h e classes of compounds not utilized include those
listed at the bottom of the figure (and the hexitols, sorbitol, and mannitol).
T h e question concerning the pentoses is based on the observation that these
compounds are used for growth b y the mutants, b u t generally only after a
prolonged lag period that is not observed with the parent strains. W e m a y
add that one of the substances not utilized in the E. call m u t a n t isolated by
Downloaded from on June 18, 2017
S. aureus ~
Published July 1, 1969
SAUL RO~MAN
Bacterial Phosphotransferase System
I55 8
Fox and Wilson (35) is the disaccharide lactose, suggesting that this disaceharide is also transported in E. coli via the phosphotransferase system ' i
Fig. 22 also shows that an Enzyme I mutant of E. coli obtained in one
laboratory (34) is unable to grow on five sugars, but can utilize others, while
a similar mutant obtained in another laboratory is unable to utilize 12 sugars
(35). These apparently conflicting results are explained by the fact that the
first mutant was shown to be "leaky" (35); in our own assays of this mutant,
5-10 % of the Enzyme I levels found in the wild type were detected in extracts
of the mutant. Thus, we may conclude that the level of Enzyme I (and presumably of HPr) required for the efficient utilization of a sugar depends on
the sugar; some sugars can be utilized at low levels of Enzyme I and HPr,
where others cannot. The range of sugars which is affected by mutations in
Enzyme I or HPr is therefore not yet established, and may have to be extended
/
(Enz. I )
,carA
/
1
/oo,
",,.
FFHa
"77v.,~SA535. /
purC///
" cysA 'CerB
his
HF,
FIOUX~ 23.
Genetic
map
of
Salmonella ~yphimurium.
(HPI)
as more "complete" mutants are obtained. To our knowledge, no deletion
mutants have yet been isolated; such mutants would not be "leaky."
Properties of an Enzyme I Mutant of Salmonella typhimurium
Both Enzyme I and HPr mutants have been obtained from S. typhimurium
(32), and the map location of the genes controlling the synthesis of these proreins has been determined (36), as shown in Fig. 23. The properties of the
HPr mutant agree with those of the Enzyme I mutant, described in detail
below.
i The special problem of lactose and galactose transport in E. coll, whether these sugars are or are
not transported via the phosphotransferase system, the role of the lVi protein in lactose transport,
and of the galactose-binding protein in galactose transport are distained elsewhere in this symposium.
Downloaded from on June 18, 2017
thr
' I~
Published July 1, 1969
~56 s
TRANSPORT
PROTEINS
T h e growth properties of the Enzyme I m u t a n t are illustrated in Fig. 24.
In these experiments, the ceils are tested in mineral m e d i u m containing the
indicated carbohydrates as the sole source of carbon, and the rate of growth is
determined. T h e values given in the figure are the time periods required for
the cell population to double, and the results show that there is no detectable
growth of the m u t a n t on nine carbohydrates (up to 10 hr), ranging from
glycerol through the disaccharides maltose and melibiose. In a rich m e d i u m
(broth), which supports growth of the mutant, it is also unable to ferment the
nine carbohydrates. In these experiments, all sugars are used at 0.2 % final
concentration in the growth medium. Similar results are obtained with
glucose at 0.5 % concentration. However, at a concentration of 1%, the
m u t a n t ceils grow slowly on glucose (generation time, 300 rain, compared
MINUTES
Pa rent
Fructose
Glycerol
Monnitol
Sorbitol
N- Acetylglucosomine
Maltose
Melibiose
Galactose
Glucose-6-P
GENERATION
TIME
Muto nt
75
120
w
95
m
130
m
80
75
80
120
90
75
70
92
75
Parent = his- 1367 r f b - 816;mutant=SBT03
Growth in mineral salts, histidine, 37°C shaker,
FIGm~ 24. Growth of Salmonella typhimurium on carbohydrates.
with 75 min in the case of the wild type). While the m u t a n t grows on three
pentoses (D-ribose, D-xylose, and L-arabinose), it requires a prolonged lag
period before growth is initiated, and with ribose, at least, the generation
time in the exponential phase is about twice that obtained with the parent
strain. Finally, the m u t a n t does grow on galactose (although at a somewhat
slower rate than the parent), and grows very well on glucose-6-P.
Analysis of crude extracts obtained from the m u t a n t (Fig. 25) shows that it
lacks detectable levels of Enzyme I; if present, the enzyme would have to be
at less than 0.5 % of the level in extracts of the parent strain. T h e figure also
shows a consistent finding that we cannot yet explain; i.e., the m u t a n t contains about three times more Enzyme II activity for methyl ~-glucoside phosphoryladon than found in the parent strain.
Genetic analysis (32, 36) of the Enzyme I m u t a n t also establishes a point
Downloaded from on June 18, 2017
Glucose
Monnose
PER
Published July 1, 1969
z57
SAUL ROSEMAN Bacterial PhosphotransferaseSystem
s
that is critical to a comparison between enzyme levels and physiological
behavior, namely that the defect results from a single mutation.
T h e two simplest explanations for the behavior of the m u t a n t are shown in
Fig. 26. T h e m u t a n t cannot utilize glucose (except at high concentrations),
but can grow as well as the parent on glucose-6-P. (We must emphasize that
glucose-6-P is transported by a specific, inducible permease system, not by
the glucose permease, reference 30). Therefore, the block in the m u t a n t must
be somewhere between the conversion of external glucose to internal glucose6-P. T h e block m a y be in transport, or it m a y be in phosphorylating glucose
after it enters the cell. T h e first will be called the transport and the second,
the phosphorylating hypothesis. In the phosphorylating hypothesis, the func-
Specific Activity in Crude Extracts*
Parent
Mutant
0.56
0.00
Enzyme lr
O. 12
0..39
HPr
O. 16
0.20
*Specific activity=Finales MeGlu-6-P formed/mg protein/50 min at 57°C
FreT.mE 25. Assay for phosphotransferase system in S. ~yphimurium.
Tronsport
GOUlIP--~G
[ / \ ] - 6 - P'==~>METABOLITESI
Phosphorylotion
out'~- in - 7 ~ G - 6 - P==I> Metab.
FIGURE26. Possible defects in
S. typhimurium mutant (I-).
tion of the phosphotransferase system is to act as a " t r a p " after glucose enters
the cell through its glucose permease.
T h e two hypotheses are readily distinguishable, since in the Uansport
mechanism, the m u t a n t should display an inability to bring the sugars it
cannot utilize into the cell, while in the phosphorylating mechanism, the rate
of entry of the sugars (and the Kin) should not be affected. (In connection with
the phosphorylating mechanism, we m a y note that the m u t a n t contains
normal levels of the specific ATP-requiring glucokinase (37) (which suggests,
but does not prove, that the defect in the m u t a n t is in transport.) O n e further
test of the two hypotheses is that if the m u t a n t can, somehow, be induced to
take up glucose, then according to the transport hypothesis it should possess
all of the machinery necessary for its utilization, and should therefore be able
Downloaded from on June 18, 2017
Enzyme I
Published July 1, 1969
TRANSPORT PROTEINS
I58 S
to grow on this sugar, whereas the phosphorylating hypothesis predicts that
it should not. The two hypotheses were tested in the following series of experiments.
T h e uptake of six labeled sugars by parent and m u t a n t cells is illustrated in
Fig. 27. In these, and in all subsequent transport experiments, uptake was
measured as quickly as possible after the addition of labeled sugar to the cell
suspension in the transport m e d i u m (7 sec), and was followed for periods up
to 10 rain. T h e figure shows that the parent strain rapidly takes up the labeled
i
i
i
I
i
i
20--
12
.I
/
Downloaded from on June 18, 2017
a
J
WILD TYPE---.
FrU.
D-
•
6
,o
20
3o
TIME
4o
50
6o
70
(seconds)
FIoum~ 27. ~4Csugar uptake.
sugars from the medium, whereas none of the sugars are taken up by the
m u t a n t at a significant rate, and in no case is uptake detected at the first
measurable time point, 7 sec. These results suggest that the defect in the
m u t a n t is in the transport process. However, the sugars employed for the
experiment shown in Fig. 27 are metabolizable, and it was therefore important
to repeat these studies with a nonmetabolizable analogue. Methyl a-Dglucopyranoside was selected for this purpose, since it is known to be transported via the glucose permease system (23). T h e results of one experiment
are shown in Fig. 28, and clearly indicate a dramatic difference between
parent and m u t a n t strains in their ability to take up the glycoside.
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
z59 s
A series of experiments of the type shown in Fig. 28 were performed with
different concentrations of methyl a-glucoside (over a 35-fold range), and
the resulting kinetic data were used to obtain the V ~ y for both parent and
mutant cells. The calculations were based on the initial time point (7 see),
which was as close as we could get to measuring the initial rate. The results
I
F"1(.9
LU 32
I
I
I
I
I
I
>- 28
r-~ 24
~E
E
20
¢rL~ 12
Q_
(./3
8
o
4
~E
::L
MUTANT
LACTATE
G R O W•N ;
"~l
, ~ 1
I
I
I
I
20 4 0 6 0 80 I 0 0 120 140
Fmtn~E 28.
side uptake.
Methyl a-gluco-
T I ME (seconds)
PARENT
1.9
MUTANT
~-MeGlu
/.~moles/g
Fmtym~ 29.
Uptake (ventry~ . . . . .
max /
0 . 0 4 (?)
dry wt /sec
Glucose permease defect in Salmonella mutant (lactate grown ceils).
are illustrated schematically in Fig. 29. There is a minimum of 50-fold difference
in the rates of entry of methyl a-glucoside between the m u t a n t and parent
strains. This ratio is considered m i n i m u m for two reasons: (a) T h e values
obtained with the m u t a n t at 7 sec were so close to the controls, formaldehydetreated ceils, that their significance must be questioned. In addition, the
values obtained with the m u t a n t at 7 see did not increase with increasing
Downloaded from on June 18, 2017
16
Published July 1, 1969
160 S
TRANSPORT PROTEINS
external concentration of methyl a-glucoside, unlike those obtained with the
parent strain. T h e estimate of 0.04 for the W.~tr~ in the m u t a n t is therefore a
guess, and represents the maximum rate of entry of the glucoside into the mutant cells. (b) T h e W,ntryma,of the glucoside into the parent cells is a minimum
value. In order to obtain initial rates of entry, it was necessary to assume that
the rate remained constant from zero time to the first measurable point, 7
sec. T h e shape of the progress curve shown in Fig. 28 indicates t h a t this assumption is probably not correct, and that the initial rate was probably m u c h
greater than observed at 7 sec. Thus, despite the linearity of the LineweaverBurke plots, we m a y conclude that the W~tr~ in the parent cell given in Fig.
28 is probably an underestimate. In any case, the results with the glycoside
I-- 2 0
-I(_9 18
I
I
I
I
60
I
80
I
I
I
L~J
>- 14
n."
a
12
rY
~D
8
n-
~
o~..a.Jo
4
2
::k
I
20
I
40
I
I
1
I 0 0 120 140
FzouP~ 30.
toside uptake.
Methyl /3-galac-
TI ME (seconds)
amply confirm those obtained with the sugars, and lead to the following
important conclusion.
A single mutation in Salmonella typhimurium results in an inability of the cell
to synthesize E n z y m e I of the phosphotransferase system. T h e physiological
consequence of this defect is an inability to grow on, or ferment, nine carbohydrates because these substances are not transported into the cell.
O n e remaining point must be emphasized, the fact that the rate of transport
is the critical parameter. In the case of the mutant, methyl a-glucoside uptake
was almost undetectable at 7 sec, but after extended periods of time, the
internal concentration of the sugar reached that of the external concentration.
Glucose is, of course, metabolized, and its internal concentration cannot be
determined. Nevertheless, a significant quantity of radioactivity was accumu-
Downloaded from on June 18, 2017
I
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
161 s
lated by the m u t a n t at 10 rain when labeled glucose was used (about 20 % of
that observed with the wild type). Thus, it appears that the problem in glucose
utilization by the m u t a n t is an inadequate rate of transport. T h e prime importance of the rate of transport is considered again later in this discussion.
If the phosphotransferase system is involved in the transport of the nine
sugars not utilized by the Enzyme I mutant, then what about the sugars that
are utilized? Are the utilizable sugars transported via a different system? To
study this question, we examined one of the utilizable sugars, galactose.
Galactose is transported by four different permease systems in E. coli (38, 39).
T h e properties of the inducible galactose transport system in Salmonella
typhimurium are identical with one of the four systems described in E. coli,
called the M G permease. This system transports methyl 3-D-galactopyranoside (which is not utilized for growth), D-galactose, and D-fucose (also not
MUTANT
PARENT
1.5
~. . . .
~-MeGoI Uptake (V
\ entry
max J
/jmoles/g
0.58
dry wf /sec
F1oum~ 3 l. Methylgalactoside (MG) permease defect in Salmonella mutant. (Grown
galactose or lactate plus fucose.)
on
used for growth), and is induced by growth of the cells on galactose, or in the
presence of D-fucose.
Transport studies were conducted with induced cells, with labeled methyl
13-galactoside as substrate, and were performed in the same manner as
described for methyl a-glucoside. One such experiment, illustrated in Fig. 30,
clearly shows a significant difference between the two types of cells. T h e
peculiar shape of the uptake curves, noted at all concentrations of methyl
/3-galactoside tested, remain to be explained.
Furthermore, and perhaps equally important, at prolonged times under a
variety of conditions, the parent strain accumulated methyl /3-galactoside,
whereas the m u t a n t could only equilibrate internal with external concentrations of the glycoside. (These studies were conducted with ceils grown on
galactose, or induced by growth on lactate plus fucose, and -4- lactate in the
transport m e d i u m as an energy source.) Thus, we conclude that the mutant,
unlike the parent, takes up galactose by facilitated diffusion.
Downloaded from on June 18, 2017
(~-MeGoI)
Published July 1, 1969
i62
s
TRANSPORT PROTEINS
A kinetic analysis of methyl/3-galactoside uptake was conducted, and the
results are illustrated in Fig. 31. T h e v~.~t,~ of uptake of the glycoside differs
by a factor of four for the parent and m u t a n t cells. Despite this difference,
the results show that unlike methyl a-glucoside, which is taken up by the
m u t a n t at a rate too slow to be measured with accuracy, the m u t a n t does
take up methyl fl-galactoside at a significant rate, by a process characterized
above as facilitated diffusion. While the facilitated diffusion process in the
m u t a n t is slower than the active process in the parent strain, it transports
galactose at a rate sufficient to support growth of the mutant.
T h e results obtained with methyl/~-galactoside in the E n z y m e I m u t a n t m a y
be explained by one of the predictions m a d e earlier in this discussion (Fig.
PARENT
MUTANT
lu I
~, HP.
~-MeGol
(Gol)
entry
Vmax
I. 5
1.9
.p- MeGaloctoside,
.£/- MeGlucpside,
/~moles/g
O. 38
0.2-0.3
dry wf /sec
FIouP~ 32. a-Methylglucoside (a-MeGlu) transport via the methylgalactoside (MG)
permease. (Galactose or lactate/fucose grown cells.)
20). T h a t is, E n z y m e I or H P r mutants should be unable to grown on, or
transport all carbohydrates utilized by the phosphotransferase system, but it
is possible in some cases for a facilitated diffusion process to occur, since the
Enzymes II are present in the membranes of such mutants. T h e rate of
facilitated diffusion is dependent on parameters such as the rate of dissociation
of the E n z y m e I I - s u g a r complex on the inside of the membrane, or more
generally, on the rate of translocation in the absence of phosphorylation.
T h e results with methyl fl-galactoside lead to one final, but important
experiment, which was performed to answer a question raised earlier. T h a t is,
if the prime function of the phosphotransferase system is in sugar transport,
then the cytoplasmic and other m e m b r a n e - b o u n d enzymes required for sugar
utilization should be intact in E n z y m e I (or HPr) mutants. Therefore, if it
Downloaded from on June 18, 2017
(Z-MeGlu
(Glu)
Published July 1, 1969
SAUL ROSEMAN BacterialPhosphotransferase System
163 s
were possible somehow to induce entry of glucose into such mutants, the cells
should be able to utilize and to grow on this sugar. Fortunately, the inducible
M G permease in E. coli is not specific for galactose, methylgalactoside, and
fucose, but also transports glucose (40). To determine whether S. typhimurium
behaves similarly, transport experiments were conducted with methyl aglucoside in m u t a n t and parent cells grown on lactate, and in a m e d i u m
containing lactate plus fucose (to induce the M G permease). As reported
above, methyl a-glucoside is not significantly transported in uninduced m u t a n t
cells (Fig. 29). However, there was a significant uptake by the induced m u t a n t
cells, and the results of the kinetic analysis are shown in Fig. 32. T h e calculated v~.~t~y for methyl a-glucoside in induced m u t a n t cells was about half
I00
90-
I
I
I
I
I
I
I
I
I|
1
8o
z,o
60-
I-hi
~l~
40
_J
Y
"~" 3 0
LLI
0
Z
,~ 2O •
~norrm / t
/+IO-'MFUCOSII
rn
<:~
F m t r ~ 33.
Y"
I0
Io"_u FUCOS_E I
,
,"
t,~ -,
,
,
,-GLU COS,EJ
I
2
5
5
6
7
4
8
9
Glucose utilization by Enzyme I mutant of Salmonella ~yphimurium mutant
induced for the MG permease. The abcissa is time in hours, while the ordinate represents
relative cell number (logt0 scale).
that obtained with methyl ~-galactoside in the same organism. Again, after
prolonged incubation, the internal concentration of methyl a-glucoside did
not exceed the external concentration, and we therefore conclude that the
M G permease-induced m u t a n t is capable of transporting methyl a-glucoside
by the process of facilitated diffusion.
Methyl a-glucoside transport in the m u t a n t induced for the M G permease
should be a measure of the ability of the m u t a n t to transport the natural
substrate, glucose. Since the m u t a n t transported galactose by facilitated
diffusion at a rate adequate to support growth, it was possible to test the
hypothesis that the m u t a n t contained all of the machinery necessary for
growth on glucose, excepting the normal transport mechanism. T h a t is, experiments could be conducted to determine whether or not the m u t a n t could
grow on glucose when it entered the cell via the inducible M G permease.
Downloaded from on June 18, 2017
~
Published July 1, 1969
i64 s
TRANSPORT
PROTEINS
Conclusions
T h e correlations outlined in Fig. 16 indicate the identity between the phosphotransferase and sugar permease systems. Detailed examination of the
properties of the S. typhimurium m u t a n t permitted us to show that the phosphotransferase system functions in the transport process according to the
predictions of a simple model (Figs. 19 and 20). Thus, we must conclude that
only one of the two hypotheses offered in Fig. 26 to explain the behavior of
the m u t a n t is correct. T h e phosphotransferase system does not act as a "sugar
trap" (phosphorylating hypothesis), but is required for sugar transport.
To our knowledge, there is no experimental evidence against this conclusion,
and as summarized above (Fig. 16), when the properties of the two systems,
phosphotransferase and permease, are compared by a variety of methods,
they agree in all respects.
DISCUSSION
T h e PEP: glycose phosphotransferase system is required for the transport of
most sugars across bacterial membranes. T h e following questions remain to be
answered: (a) Are all sugars transported in all bacteria by the phosphotransferase system or some modification of it? (b) Are other solutes transported by
this system, or some modification of it? (c) Do other organisms contain the
phosphotransferase system? (d) How is the phosphotransferase system controlled, so that sugars are transported only at rates that permit their utilization,
and do not result in the accumulation of excessive quantities of the sugar or
Downloaded from on June 18, 2017
T h e results of two experiments along these lines are shown in Fig. 33. In
one case, the m u t a n t was grown on galactose in a mineral medium, washed,
and transferred to a glucose medium. Growth commenced as indicated in
in the figure, without a lag, and at a rapid rate, but the rate decreased with
time. T h e reduction in growth rate is explained by dilution of the M G permease, by growth under noninducing conditions.
In the second type of experiment, the m u t a n t was first grown on nutrient
broth and therefore did not contain the induced M G permease. After washing,
the cells were transferred to mineral m e d i u m containing glucose, or fucose, or
both sugars. After a lag period of the type normally observed in inducing
systems, the m u t a n t grew in the m e d i u m containing both sugars, and at a
slow, but exponential rate.
Thus, from these experiments it can be seen that the internal machinery for
glucose utilization is intact, and that glucose can be utilized for growth provided that it can enter the m u t a n t cell. T h e only defect in the m u t a n t is in the
transport process, and the majoi function of the phosphotransferase system
must be in transport of carbohydrates.
Published July 1, 1969
SAUL ROSEMAN BacterialPhosphotransferaseSystem
~65 s
sugar phosphate? (e) In molecular terms, what is the precise mechanism by
which the phosphotransferase system mediates transport?
Some speculation can be offered in an attempt to answer the last question.
First, we must note the possible difference in binding sites of the complex
called Enzyme II on the inside and the outside of the membrane. As illustrated
in Fig. 34, the outside should only bind sugar, while the inside m a y bind one
Out
I
Membrane
/n
I
T -==L. Mg
++- P
[IL
Sugar
f Sugar
Sugar :.:~>:.:.>>>,
HPR
P-HPR
Enzyme T
Factor L[T
I
EnzymeII Complex=2 Proteins + Lipid
FiougE 34. Possible binding
sites: Enzyme II.
ouT , i,HPr
S+
+S-P
'
PEP
, ~ ~ HPr
1I
q[T+s
l~muz~ 35.
1[
S HPr
]1 ÷ Pi
Sugar transport via the phosphotransferase system.
or more of m a n y components. These differences could well lead to important
conformational changes in Enzyme II, changes that m a y actually determine
whether or not a sugar can be transported by this system. Such conformational
changes would lead to transport in a manner analogous to the "translocase"
of Mitchell (22), except that they m a y even be more complex.
A simple scheme for explaining the three processes of interest, facilitated
diffusion, active transport, and group translocation of sugars is shown in Fig.
35. While the model shown in the figure m a y resemble science fiction, it does
offer a working hypothesis. T h e sugar on the outside is b o u n d to its site on
Downloaded from on June 18, 2017
I
<':<~
Published July 1, 1969
~66 s
TRANSPORT
PROTEINS
POSSIBLE P H Y S I O L O G I C A L I M P L I C A T I O N S
Is the rate of sugar transport independent of, or related to the rates of other
important physiological processes such as glycolysis and growth? How is the
rate of sugar transport controlled?
Since there is almost no information available to answer these questions,
the following discussion can only be speculative, teleological, and will involve
many assumptions. Nevertheless, these considerations may be of value since
they will lead to hypotheses that can be examined experimentally.
A bacterial cell must adapt to widely divergent, and sometimes rapidly
changing environmental conditions. If the rate of sugar transport cannot be
varied to meet these changing conditions, the result can be disastrous. For
example, if the rate of transport is too low, growth will be limited or negligible.
If the rate is too high, energy will be wasted and potentially toxic substances
will accumulate; sugar phosphates are apparently toxic at high concentrations.
Thus, under ideal conditions, the rate of sugar transport will be regulated so
that sugar is transported into the cell only as it can be metabolized.
To meet these requirements, the transport process and its regulatory
mechanisms should exhibit the following characteristics:
1. In actively growing cells, sugar transport should be closely linked to and
controlled by anaerobic glycolysis, so that the rates of both processes are
interdependent. Stationary phase cells, which accumulate storage products
such as glycogen, will be considered separately.
Downloaded from on June 18, 2017
Enzyme II. A conformational change occurs which brings the sugar to the
inside, still bound to Enzyme II. If the complex dissociates at a rapid rate,
this process would explain facilitated diffusion (lower left of Fig. 35), a
process that does occur in a few cases of bacterial mutants lacking Enzyme I
or HPr. If the binding is tight, no appreciable dissociation will occur. Enzyme
II also has a binding site for HPr, and when this protein is bound, the complex
on the inside of the membrane consists of Enzyme II-sugar-HPr. Phosphorylation of the complex by Enzyme I yields the phospho-complex. If the phosphoryl group is transferred to the sugar, then the important process of group
translocation is explained (top right of Fig. 35). If the complex dissociates with
concomitant release of HPr, Pi, and sugar, then the process of active transport
is explained. The latter process resembles in many respects the model for
active transport proposed by Kepes (23).
Clearly, we are only at the earliest stages in reaching a full understanding
of sugar transport in bacteria. We must eventually comprehend precisely how
the phosphotransferase system mediates this process, and finally, how the
Enzyme II complex is inserted into the membrane to become fully functional.
However, it is equally clear that experiments can now be conducted that may
provide answers to these questions.
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
x67 s
2. In actively growing cells, the rate of sugar transport should be inversely
related to the concentration of A T P available to the cell, but directly related
to the rate at which A T P is utilized. In other words, when the quantity of
A T P is high, and greatly exceeds the need for biosynthetic reactions (other
than for the synthesis of storage products), then sugar transport should stop;
however, if A T P is being used at a rapid rate, then the rate of sugar transport
should be maximal to supply the necessary fuel for A T P synthesis.
3. T h e energy used for sugar transport should be conserved where possible,
particularly under anaerobic conditions.
4. W h e n one or more steps are required to convert the transported sugar to
an intermediate in the glycolytic pathway, then the rate of transport should
not exceed the rate of conversion to the glycolytic intermediate.
5. In a m e d i u m containing more than one sugar, all except one should be
excluded to prevent the synthesis of extraneous, inducible proteins. W h e n the
~OUT
Enzymes
Phosphotransferase
Rate
Vp
orSIN
Catabolic
vc
FRUor6-P
Triose-P
~ A D P,Pi
TP
Glyca
lytic
(anaerobic)
vg
Fxom~z36. Feedback control
of transport via glycloysis
(group translocation; active
transport).
first sugar has been depleted from the medium, then a second should be
admitted to the cell (diauxic growth).
6. W h e n the uptake and catabolism of a sugar depends upon the synthesis
of inducible proteins, and the sugar is the sole source of energy available to
the cell, then transport of the sugar should take place, even at a low rate, to
permit induction of the necessary proteins.
T h e process of sugar transport, mediated by the phosphotransferase system,
exhibits m a n y of the characteristics enumerated above. T h e other properties
of the sugar transport process can be explained by postulating certain features
of the phosphotransferase s)stem that remain to be proved experimentally.
In discussing these properties, frequent reference will be m a d e to Fig. 36,
which shows the relationship between sugar transport and anaerobic glycolysis.
T h e figure refers to the primary events that occur after a sugar is transported; i.e., conversion of the sugar to glycolytic intermediates. Reference is
Downloaded from on June 18, 2017
~
Published July 1, 1969
I68 s
TRANSPORT
PROTEINS
Conservation of Metabolic Energy
As shown by Kepes (43), a molecule of A T P or its equivalent (i.e., PEP) is
required to transport a molecule of sugar against a gradient. In metabolic
systems that operate aerobically, where sugars are converted to CO2, the
expenditure of a molecule of A T P to bring the sugar into the cell is not
particularly wasteful in view of the large yield of A T P derived from the sugar.
However, under anerobic conditions, the situation is quite different; If a
molecule of glucose is transported into the cell at the expense of a molecule
of ATP, and the transport product is not glucose-6-P, but glucose per se, then
the combined processes of transport and anaerobic glycolysis (either to lactate
or to ethanol plus CO2) would yield 1 molecule of ATP. T h e decided advantage of group translocation by the phosphotransferase system, is that the
energy used for transport is not wasted. In the case of glucose, the transport
product is glucose-6-P, and the combined processes of transport and glycolysis
give rise to 2 moles of A T P per mole of sugar taken up from the medium.
Similar considerations apply to a number of other sugars, including mannose,
fructose, mannitol, N-acetylglucosamine, etc. In each case, the product of
transport is the phosphosugar, and the energy used for transport is conserved.
The idea that sugars should be converted to their phosphate derivatives
during transport was postulated many years ago, and received considerable
attention in early work on sugar transport (20). Unfortunately, the hypothesis
Downloaded from on June 18, 2017
not made to three other processes, which will be considered separately; exit of
the sugar from the cell, conversion to storage products, and utilization for
synthesis of polymers such as capsular polysaccharides and cell walls. Under
conditions of active growth, the bulk of the transported sugar is used for fuel,
a smaller quantity for the synthesis of cell materials, and only traces for the
synthesis of storage products (41, 42).
Two types of control over the phosphotransferase systems are postulated,
and they are designated primary and secondary. T h e primary controls will
include HPr, phospho-HPr, the external sugar, and the product of sugar
transport per se (either the internal phosphorylated or free sugar). Secondary
controls will consist of genetic regulation, the catabolic and glycolytic enzymes,
and the products obtained from glycolysis, PEP, and ATP. T h e more obvious
regulatory mechanisms involve the secondary controls, and these will be
considered first. Genetic control has been discussed at length; for example,
most Enzymes II are inducible, and the concentration of Enzyme II for a
particular sugar will control its rate of transport.
Transport in the following discussion means either group translocation or
active transport, unless specific reference is made to facilitated diffusion, a
process that is only of occasional significance in bacteria.
Published July 1, 1969
SAUL ROSEMAN
Bacterial Phosphotransferase System
i6 9 s
Downloaded from on June 18, 2017
was discarded when the transport process was found to operate independently
of hexokinase.
Some recent examples where the phosphosugar is required for the first step
in catabolism, and is formed during transport, are particularly interesting. An
E. coli /3-glucosidase requires the corresponding 6-phosphate ester of the
glucoside, not the free sugar; the 6-phosphate ester is the product of the transport process (35). Similarly, the/3-galactosidase of S. aureus hydrolyzes lactoseP, not lactose, and it is lactose-P that is the product of the transport process
(11).
In a teleological sense, the phosphorylated sugars should be the products
of transport of many, but not of all sugars. For example, E. coli fl-galactosidase
requires lactose, not lactose-P, and lactose is the final product of transport in
this organism. Similarly the phosphotransferase system converts galactose to
galactose-6-P, but there is no known metabolic pathway for this sugar-P;
that is, there is no known mutase that converts galactose-6-P to the desired
metabolite, galactose-l-P. (Galactose-l-P is formed from galactose by the
widely distributed galactokinase and ATP.) Thus, it appears reasonable that
the final product of galactose transport is the free sugar (44, 45), despite the
fact that galactose-6-P has been reported in one strain (29).
From these considerations, we conclude that the energy required for transport is conserved where possible, and the process operates by group translocation, but where the desired transport product is the free sugar, then the process
of active transport is invoked.
The same point of view may be extended to the facilitated diffusion process.
As considered in detail below, the energy requiring transport processes, group
translocation and active transport, may be under strict control by various
regulatory mechanisms, particularly by coupling to glycolysis. However,
facilitated diffusion requires no metabolic energy, and since the uncontrolled
entry of sugars into the cell may inhibit metabolic processes, rapid facilitated
diffusion could be disadvantageous. The results obtained with the S. typhirnuriurn Enzyme I m u t a n t indicate that the rate of facilitated (or possibly
passive) diffusion is very slow with most sugars. For example, the m u t a n t was
unable to grow on 0.2 % and 0.5 % glucose, but did grow slowly at 1% concentration. Similarly, the rate of entry of metabolizable sugars was extremely
low in the m u t a n t (Fig. 27), and the rate of entry of the nonmetabolizable
analogue, methyl a-glucoside, was too low to be accurately measured.
However, the glucoside did enter the cell, since at prolonged incubation times,
the internal concentration finally reached the external concentration.
Galactose, however, was an exception. T h e m u t a n t grew on galactose at
almost normal rates, and detailed kinetic studies showed that the facilitated
diffusion rate in the m u t a n t (of methyl ~-galactoside) was about 25 % of the
Published July 1, 1969
I7O s
TRANSPORT
PROTEINS
Exit Process
One control mechanism that could be used to maintain internal concentrations of sugar or sugar-P at utilizable levels, when the external m e d i u m contains excessive quantities of the sugar, is the exit process, which would simply
move excess carbohydrate out of the cell.
M a n y studies have been conducted on the exit process in bacteria, and
some of these have been reviewed (23). In the case of galactose, for example,
entry and exit m a y be mediated by different systems (44, 45), although a
possible complicating factor with galactose is that it is transported by several
permeases (39). T h e lactose system in E. coli has been studied extensively (23,
46), and here, and in general (23), it appears likely that exit and entry of each
sugar is mediated by one system.
An exit mechanism of particular interest is called counter-flow (or countertransport), which is not energy-requiring, and is m e a n t to describe the process
where a molecule of sugar external to the cell exchanges with a molecule inside
the cell. T h e two sugars can be different, and the internal compound m a y even
Downloaded from on June 18, 2017
rate of transport in the parent strain. This result can be rationalized by the
arguments used above.
T h e desired end product of galactose transport (and lactose transport in E.
coli) is the free sugar. If 1 mole of PEP is required to transport this monosaccharide, the combined processes of transport and glycolysis under anaerobic
conditions would yield 1 mole of ATP. However, when the rate of facilitated
diffusion is adequate to supply the sugar to the cell, the net yield is 2 moles
of ATP. Thus, it is possible to understand why galactose, unlike the other
monosaccharides, can be taken up at a significant rate by facilitated diffusion.
If facilitated diffusion operates at an adequate rate, why then is there
another, energy-requiring process (active transport) available to the cell?
Perhaps the answer lies in the fact that the cell must be capable of growth at
low, as well as at high concentrations of sugar in the medium. A m i n i m u m
internal concentration of sugar is required to support the m i n i m u m rate of
glycolysis necessary for growth. At low external sugar concentrations, active
transport, unlike facilitated diffusion, can maintain this m i n i m u m internal
concentration of sugar. These suggestions lead, however, to an additional
problem.
For each bacterial species, there are apparently two classes of sugars, one
which is transported at a significant rate by facilitated diffusion, and one
which is not. Regulation of the transport rate of the latter group can be accomplished by several mechanisms, such as the level of PEP, and these controls
are considered in detail below. For the class of sugars where two mechanisms
are available, facilitated diffusion and active transport, what are the controls
which determine whether one or the other process is invoked?
Published July 1, 1969
SAUL ROSEMAN
Bacterial Phosphotransferase System
I7! s
Coupling of Transport to Glycolysis; Pivotal Role of PEP
Phosphoenolpyruvate can be synthesized by several reactions, but we will
assume that the major pathway is anaerobic glycolysis in cells actively growing
on carbohydrates. O n this basis, the end product of sugar transport followed
by glycolysis is PEP (Fig. 36), and the end product, PEP, is required for the
transport of more sugar. Sugar transport is thus under feedback control by
its product. In biochemical systems, this type of feedback regulation is unusual;
most feedback processes involve inhibition; i.e., the end-product inhibits the
first reaction. Here, the end product is required for the first reaction.
PEP, rather than ATP, is a suitable candidate for the driving force behind
sugar transport for several reasons. First, PEP participates in relatively few
reactions, and the internal pool of PEP may be more easily regulated than that
of ATP. Second, the rate of PEP synthesis will depend on the rate of anaerobic
giycolysis, which should be relatively independent of fluctuations in the degree
of aerobiosis; the rate of A T P synthesis could fluctuate considerably with
Shifts in aerobiosis. Third, it is important from considerations raised earlier
that the rate of sugar transport not be completely independent of the rate of
A T P synthesis. Since anaerobic glycolysis is controlled by aerobic oxidation,
the necessary balance for PEP synthesis can be achieved as discussed below.
To examine these points, Fig 36 shows the rates of three processes: vp (sugar
transport); vo (the rate of conversion of the transport product to a glycolytic
intermediate by enzymes such as glycosidases, isomerases, etc.); and v0 (the
rate of anaerobic glycolysis).
In the growing ceil, steady states are achieved for at least brief periods of
Downloaded from on June 18, 2017
be the sugar phosphate (30, 31). Under special circumstances, this type of
exit may be physiologically significant; for example, when the level of an
inducible permease is negligible, counterflow may be invoked to bring the
inducer into the cell via a different permease.
Despite the demonstrated ability of bacterial cells to transport sugars from
inside to outside, we do not believe the exit process to be of importance in
growing cultures. Lactose is utilized as rapidly as it is taken up from the
m e d i u m (23, 43). From our own experiments, we believe this is also true for
glucose. Furthermore, nonmetabolizable analogues such as T M G and methyl
a-glucoside are lost from cells at very low rates when the cells are placed in
sugar-free media (provided that the cells can supply the necessary energy to
maintain the gradient). Finally, under anaerobic conditions, the exit process
could lead to cessation of growth, since energy would be wasted by transporting a sugar into the cell at the expense of a mole of PEP, and then permitting
it to leave.
If the exit process is inconsequential during normal growth, how is the
internal sugar concentration maintained at the desired level?
Published July 1, 1969
z72 s
TRANSPORT
PROTEINS
Downloaded from on June 18, 2017
time (41, 47). Under these conditions the concentration of metabolites (in
Fig. 36) remains relatively constant, and vp = vo = vg the transported sugar
will be used at the rate at which it is brought into the cell. Since the rate of
synthesis of PEP (vg) depends upon the rate of sugar transport (vp), and conversely, the rate of sugar transport depends upon the rate of synthesis of PEP,
it is seen that the transport and glycolytic processes are mutually self-regulating.
Under o p t i m u m conditions, we may presume that the rate-controlling factor
is sometimes V e ~ of the sugar, and that vo and v0 are capable of handling
sugar as fast as it can be transported; for example, the lactose operon in E. coli
can synthesize 50 times more/3-galactosidase than is required to hydrolyze
the m a x i m u m quantity of lactose that can be transported.
In addition to the limiting situation determined by V ~ y, the other limit
would result when anaerobic glycolysis stops. The rate of glycolysis is determined by a number of factors; only one will be examined here. When cells are
in a medium containing adequate quantities of inorganic phosphate, anaerobic
glycolysis is controlled by the availability of ADP. In cells respiring at a rapid
rate, but where the demand for A T P is limited, then most, if not all of the
available adenosine nucleotide will be in the form of ATP. The lack of ADP
brings glycolysis to a halt. Thus, despite m a x i m u m levels of ATP, PEP is not
being synthesized, and sugar transport stops. In fact, Kepes (43) has called
attention to the lack of correlation between sugar transport rates and the
levels of intracellular ATP.
These theoretical considerations explain certain experimental observations.
At concentrations of azide and dinitrophenol that should uncouple oxidative
phosphorylation, the transport rate of methyl a-glucoside is greatly stimulated,
rather than inhibited (48). The explanation for this observation is that uncoupling of oxidative phosphorylation is known to increase the rate of formation
of PEP (49). T h a t PEP plays a pivotal role in the uptake process is also supported by the fact that under the conditions where azide stimulates methyl
a-glucoside transport, the well known anaerobic glycolysis inhibitor, iodoacetate, markedly depresses the rate (48). Similar considerations apply to the
results of experiments where the cells are supplied with rapidly metabolizable
compounds during methyl a-glucoside uptake (43, 50, 51); these compounds
are found to inhibit uptake. However, the addition of azide to the medium,
which forces the cell to utilize the rapidly metabolized substances exclusively
by anaerobic glycolysis, markedly stimulates methyl a-glucoside transport.
T h e authors state that "accumulation is greater when the cells depend for
their energy supply on endogenous respiration or on the anaerobic fermentation of an exogeneous substrate rather than on the active oxidation of such
a substrate in the absence of poisons" (50).
Thus we may state that evidence exists to support the contention that sugar
transport is linked to anaerobic glycolysis, through the requirement for PEP,
Published July 1, 1969
SAUL ROSEMAN BacterialPhosphotransferaseSystem
x73 s
Other Metabolic Pathways; Uncoupling of Transport and Glycolysis
Transported sugar has thus far been assumed to travel the pathway shown in
Fig. 36, where transport and glycolysis are coupled. However, if the bulk of
the sugar that is transported into the cell escapes metabolism via this pathway,
then the simplification shown in Fig. 36 is invalid. O n e such possible fate of
transported sugar, the exit reaction, has already been considered, and coneluded to be of little physiological significance under the usual conditions of
growth. Other possible metabolic fates that await transported sugar are
catabolism by pathways other than anaerobic glycolysis, the synthesis of
storage products such as glycogen, and the synthesis of polysaccharides such as
those found in the cell wall, capsule, and lipopolysaccharides.
T h e major catabolic pathway for sugar utilization in most strains of bacteria
appears to be glycolysis through the Embden-Meyerhof-Parnas pathway
Downloaded from on June 18, 2017
and that regulation of glycolysis by any of the available mechanisms, such
as the ratio of A D P to ATP, will in turn control the rate of sugar transport.
T h e discussion has thus far centered on two of the three reactions shown in
Fig. 36, transport and glycolysis. Is one of these processes always rate limiting,
or can the rate-limiting step sometimes be the catabolic pathway leading to
glycolysis, which determines vo? I n this connection, it is interesting to note that
the parent strain of S. typhimurium grows on 16 carbohydrates, but that the
growth rate is o p t i m u m on glucose, glucose-6-P, galactose, and sorbitol (32).
T h e rate of growth is less, frequently substantially less, on the other 12 carbohydrates. I n some of these cases, at least, it appears possible that the ratelimiting step m a y be vc, and not v~.
Obviously, a m i n i m u m rate of glycolysis must be achieved to support
growth, and according to Fig. 36, this means that a m i n i m u m rate of transport and catabolism must likewise be achieved for growth. This concept explains the growth properties of parent and m u t a n t strains of S. typhimurium on
different sugars. T h a t is, sugars that are transported at a high rate in the
parent strain (glucose and galactose) continue to support growth of the
m u t a n t even when the transport rate is I0-15 % of that obtained in the parent
strain (32). O n the other hand, growth of the m u t a n t is not detected on a
sugar that is poorly utilized by the parent, such as mannose, although preliminary results indicate that mannose is taken up by the m u t a n t at 25 % of
the rate obtained with the parent. Presumably, reduction in the transport rate
of mannose cannot be tolerated because this decrease in the transport rate
depresses the rate of glycolysis below the required m i n i m u m value necessary
for growth.
T h e emphasis in this discussion has been that the rate of sugar transport is
regulated to correspond to the rate at which it is used, but are the transport
and glycolytic processes always coupled?
Published July 1, 1969
~74 s
TRANSPORT
PROTEINS
Downloaded from on June 18, 2017
(anaerobic glycolysis). However, two other degradative systems are known, the
pentose phosphate pathway of Warburg-Dickens-Horecker-Racker, and the
2-keto-3-deoxy-hexonic pathway of Entner-Doudoroff-Wood; in some bacteria, one or the other of these two pathways is the major catabolic system for
degrading sugars. The products of the latter two pathways, however, are
further catabolized through the glycolytic pathway, and these degradative
pathways are therefore considered to be the second step in Fig. 36, which
determines rate vc. On this basis, they need be of no fur.ther concern to this
discussion.
When a carbohydrate is the sole source of carbon available to a growing
cell, almost all constituents of the cell are synthesized from cleavage products
of the sugar obtained by anaerobic glycolysis, aerobic respiration, the degradative pathways discussed above, etc. In these cases, therefore, the transported
carbohydrate follows the path shown in Fig. 36, and transport and glycolysis
are coupled. However, cell polysaccharides are not synthesized by this route.
Numerous isotope experiments have demonstrated that sugars are incorporated directly (i.e., without fragmentation) into polysaccharides, and here,
the transport process must provide sugar above that required for glycolysis.
In quantitative terms, what fraction of the transported sugar is used for the
biosynthesis of polysaccharides?
When E. coli grows on glucose as the sole source of carbon, about 75 % of
the sugar undergoes glycolysis, and about 25 % is converted to cellular
material (41). About 10 % of the latter consists of polysaccharides (cell wall,
lipopolysaccharides, etc.) (52). Thus, abou,~ 2.5 % of the total sugar used by
the cell for growth is incorporated into these structural polysaccharides. It is
therefore reasonable to conclude that in the growing cell only negligible
quantities oi sugar escape the pathway shown in Fig. 36.
Stationary phase cells, and cells that normally produce capsules, can be
made to synthesize much larger quantities of polysaccharide. When some
strains of capsule-producing bacteria are grown under conditions where
nitrogen is limiting, increased quantities of extracellular polysaccharide are
formed. Stationary phase cells, maintained in a m e d i u m rich in sugar, can
accumulate storage products; glycogen, for example can be accumulated
up to 20 % of the cell mass. The formation of large quantities of polysaccharides at first suggests that a substantial quantity of transported sugar
escapes the pathway shown in Fig. 36, and at least this portion is not coupled
to the rate of glycolysis. This conclusion may prove correct, but further consideration suggests that it may not. For example, in the case cited above, if
20 % of the cell mass of E. coli is glycogen, and I 0 % consists of the structural
polysaccharides, the total polysaccharide synthesized from glucose transported by the cell amounts to 7.5 %. Therefore, even in cells that grow to the
Published July 1, 1969
SAUL ROSEMAN
Bacterial Phosphotransferase System
I75 s
Some Possible Primary Controls over Sugar Transport
As indicated earlier, primary controls are those involved directly in the sugar
transport process, rather than those affected by subsequent metabolic reactions.
The need for such controls may be illustrated by two examples: (a) Diauxic
growth means utilization of one of two (or more) sugars until the medium is
depleted of this compound, followed by induction of the required enzymes and
the permease for utilization of a second sugar. Diauxie is a complex phenomenon, involving regulation of induction, catabolite repression, etc. Nevertheless,
Downloaded from on June 18, 2017
stationary phase, the bulk of the transported sugar traverses the route shown
in Fig. 36.
Is there any connection between the regulatory mechanisms for glycogen
synthesis, sugar transport, and glycolysis in the resting cell, or are these
processes uncoupled? PEP has been implicated as a prime factor in sugar transport, and it is therefore of particular interest that the E. coli enzyme which
catalyzes the synthesis of the glycogen precursor, ADP-glucose, is activated
20- to 40-fold by three products of anaerobic glycolysis, fructose diphosphate,
glyceraldehyde-3-P, and PEP (53). In medium containing sugar, stationary
phase cells synthesize lipid, protein, nucleic acid, etc. at negligible rates, but
can transport sugar, and degrade it by glycolysis. The two processes can
continue until glycolysis stops because of lack of ADP. However, if the ADPglucose pyrophosphorylase is activated by the resulting high levels of hexose-P
and PEP, the three processes, transport, glycolysis, and glycogen synthesis can
be coupled. This follows from the fact that the net reaction for glucose transport (by group translocation) and glycogen synthesis is: glucose + [glycogen]
+ ATP + PEP --. glucosyl-[glycogen] + ADP + pyruvate + PPi. The
incorporation of 1 mole of glucose into glycogen therefore requires 2 moles of
energy-rich compounds (PEP and ATP), and these 2 moles can be provided
by the transport of an additional mole of glucose, followed by glycolysis. Under
optimum anaerobic conditions therefore, 2 moles of glucose would be transported into the cell, one would be incorporated into glycogen, and the other
undergo glycolysis. The key point is not the energy balance, however, but
rather that glycogen synthesis requires the conversion of ATP to ADP. As
discussed earlier, it is precisely this step that may be required to permit the
transport of another mole of sugar (by stimulating glycolysis) which results in
the formation of the required compound, PEP. Thus, we can see the possibility
of three interlocking systems, transport, glycolysis, and glycogen synthesis,
where the rate of one process regulates the rates of the other two. Similar
considerations would apply to the synthesis of other bacterial polysaccharides,
such as the capsular materials.
Published July 1, 1969
i76 s
TRANSPORT
PROTEINS
COMPETITION FROM WITHOUT AND INHIBITION FROM WITHIN
T h a t sugars
can compete with each other for a transport site is a well-documented
phenomenon. T h e example of glucose and galactose was cited above. (It is
interesting to note that glucose is a potent inhibitor of galactose phosphorylation in the phosphotransferase enzyme system, while galactose inhibits T M G
phosphorylation.) In this connection, we m a y again call attention to the fact
that E n z y m e II consists of two proteins, A and B, and that fraction II-A
appears to contain different, sugar-specific proteins. T h e conformational
change shown in Fig. 35 m a y be envisioned as a translocation of protein, At,
from outer to inner side of the membrane, while it remains complexed to
protein B, and carries with it a molecule of sugar, $1. This type of mechanism
is completely analogous to the conventional sugar "carrier" hypothesis,
where the carrier shuttles from one side of the m e m b r a n e to the other. Two
possible modes of competition from without are: (a) Two sugars, $1 and $2,
compete for the specific carrier proteins, A~. (b) Protein B also has the capacity
of binding sugars, but it is nonspecific (and m a y also have a low affinity
relative to A). Following the translocation step, where S1-A~ moves from the
outer to the inner edge of the membrane, B is exposed to the external medium.
If B binds S= rather than $1, the return of A1 from inside to outside m a y be
prevented, and thus prevent further transport of sugar S1. Alternatively,
Downloaded from on June 18, 2017
in at least one case, where the m e d i u m contains glucose and galactose,
diauxie is explained by inhibition of sugar transport; glucose inhibits entry
of galactose, and the enzymes required for galactose utilization cannot be
induced (25). (b) W h e n nonmetabolizable substances are used in transport
experiments (Fig. 28), the rate of transport decreases substantially after the
first few seconds, despite the fact that energy is available to the cell. O n e
explanation for this phenomenon is that the internal sugar leaves the cell as
its concentration rises, and the rate of net accumulation therefore declines.
However, this explanation appears improbable since the internal concentration of the sugar is very low when the transport rate decreases, far below the
concentration of sugar in the medium.
Primary controls over sugar transport are easily visualized (by the biochemist
who is willing to substitute imagination for unavailable fact), particularly
because of the unique topographical location of E n z y m e II. Unlike soluble
enzymes which are exposed on all sides to the same solutes, E n z y m e II is
exposed on one side to the constituents of the growth or transport medium,
and on the other to the constituents of the cytoplasm. An attempt to depict
the asymmetry of Enzyme II is offered in Fig. 34, and a mechanism for transport based on the asymmetry is shown in Fig. 35. Obviously, since we have no
precise information, variations of the mechanism shown in Fig. 35 can be
offered to explain all of the phenomena involved in sugar transport. Two such
possible regulatory mechanisms will be considered.
Published July 1, 1969
SAUL ROSEMAN
Bacterial Phosphotransferase@stem
277 s
REGULATION BY HPR, PHOSPHO-HPR, AND ORDER OF BINDING
The critical
event in the mechanisms for sugar transport depicted in Fig. 35 is the conformational change, or translocation, that brings the sugar from the outside
to the inside of the ce11. In the sequence shown in the figure, which m a y be
designated HPr-fndependent translocation,H P r binds to the complex after the
translocation step, and is then phosphorylated. Another important possibility
is where phospho-HPr binds to Enzyme 11, and the translocation step occurs
as a resultof this binding; this mechanism m a y be designated phospho-HPrdependent translocatfon.In the latter case, H P r could conceivably prevent
transport by competing with phospho-HPr for the binding siteon Enzyme II.
Reference was made earlier to the fact that there are two classes of sugars
for each strain of bacteria, a class that can be taken up from the m e d i u m by
the process of facilitateddiffusion at a significantrate, and a class that cannot.
The mechanism shown in Fig. 35 explains the properties of the first class,
while the phospho-HPr-dependent mechanism explains the properties of the
second class. Furthermore, the potential inhibitory activity of H P r in the
phospho-HPr-dependent mechanism explains w h y mutants which have a
limited ability to phosphorylate H P r (Enzyme 1 "leaky" mutants) can
utilize some, but not all sugars; the degree of inhibiton would depend on the
relative affinitiesof H P r and phospho-HPr for the Enzyme II complex.
Finally, the two mechanisms considered above explain w h y a single
inhibitor, such as the oxidative phosphorylation uncoupler, sodium azide, can
have opposite effects on the transport rates of two different sugars. Azide
stimulates methyl o~-glucoside uptake (as described earlier), but inhibits the
active transport of T M G . If methyl ~-glucoside is taken up by the phospho-
Downloaded from on June 18, 2017
specific protein A1 m a y transport competing sugar S,, if this sugar first binds
to B. T h e sequence would be: $2 + B - A 1 ~ S ~ - B - A 1 ~ S , - A 1 - B
B -A1 -$2.
This model also explains counter-flow. A sugar on the inside of the cell,
S~, binds to its specific site, A~, while the external sugar, $2, binds to the nonspecific site, B. Exchange of one sugar for the other results in counter-flow.
W h e n the internal sugar is the phosphate ester, the protein complex A-B
would also have to act as a transphosphorylase during the reaction.
Finally, the same model can be used to explain inhibition from within, a
process that could prevent the accumulation of potentially toxic sugars and
sugar phosphates, and that could result in a very rapid decline in the rate of
transport of substances such as methyl a-glucoside (cited above). Inhibition
from within is visualized to occur as indicated above, except that the same
sugar is on each side of the m e m b r a n e (it can be the sugar phosphate on the
inside). T h e E n z y m e II complex would then be: S ~ - A s - B - $ 1 . Either translocation is prevented, or counter-flow occurs. In either case, the process would
be inhibitory to the net flow of sugar from outside to inside.
Published July 1, 1969
I78 S
TRANSPORT PROTEINS
Conclusions
An attempt has been made to explain the behavior of ceils under widely
divergent conditions on the basis of some known properties of the phosphotransferase system, and some that are speculative. As indicated earlier, however, many of these speculations can be tested. For instance, the isolation of
components of the phosphotransferase system in homogeneous form will permit direct determination of the various interactions that have been postulated.
With these quantitative values in hand, definite conclusions on the validity of
these hypotheses may be obtained.
Reference has been made to the contributions of the author's associates and he would llke to express
his thanks for their many helpful discussions during the preparation of this manuscript. We are particularly grateful to Dr. Harry Schachter for his advice on the kinetics of transport.
Contribution 547 of the McCollum-Pratt Institute. These studies were supported by Grants A M
00512 and A M 09851 from the National Institute of Arthritis and Metabolic Diseases of the National
Institutes of Health.
REFERENCES
1. ROSEMAN, S. 1968. Biosynthesis of glycoproteins, gangliosides a n d related substances. In
Biochemistry of Glycoproteins a n d Related Substances, Proceedings of the 4 t h I n t e r -
Downloaded from on June 18, 2017
HPr--dependent mechanism, and T M G by the mechanism shown in Fig. 35,
then precisely these results are expected. The inhibitor greatly stimulates the
production of PEP, which may convert most of the cellular H P r to phosphoHPr. The increased availability of phospho-HPr would increase methyl aglucoside uptake. However, the decreased quantity of H P r would inhibit
active transport of T M G , as can be deduced from Fig. 35, although it would
not affect facilitated diffusion of T M G . In fact, azide does not inhibit facilitated diffusion of T M G but inhibits its retention by the cell (46).
We have speculated that there may be two (or more) mechanisms for sugar
transport, HPr-independent and phospho-HPr-dependent translocafion; possibly neither one of these mechanisms is correct. However, the important
question is whether a single system, such as the phosphotransferase system,
possess sufficient inherent flexibility to explain the observed variations in
sugar transport systems; a single sugar may be transported differently by
different organisms (lactose by E. coli and S. aureus), while a single organism,
such as E,. coli, transports different sugars by more than one process (facilitated
diffusion, group translocation, active transport). The two mechanisms offered
above explain these phenomena. For example, what at first appears to be a
trivial change, i.e., the order of interaction of the constituents of the phosphotransferase system, results in a qualitative change in the transport process
(facilitated diffusion can or cannot occur). We therefore conclude that the
phosphotransferase system is flexible enough to act as a general mediator of
sugar transport in bacterial cells.
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
national Conference on Cystic Fibrosis of the Pancreas (Mucoviscidosis), Part II. E.
Rossi and E. Stoll, editors. S. Karger, Basel. 244.
COMB, D., and S. RosnM.~. 1960. J. Biol. Chem. 235:2529.
KtrNDIO, W., S. GHOsH, and S. RosEM~a~. 1966. J. Biol. Chem. 241:5619.
KtrNnxo, W., S. GHOSH, and S. ROSnMAN. 1964. Proc. Nat. Acad. Sci. U.S.A. 52:1067.
Kt~DIG, W., and S. ROSEMAN. 1966. A phosphoenolpyruvate hexose phosphotransferase
system from Escherichia coli. In Methods in Enzymology. W. Wood, editor. Academic
Press, New York. IX:396.
K~DIO, W., F. D. KtrNDIO, B. E. ANDERSON,and S. ROSEMAN. 1965. Fed. Proc. 24:658.
ANDERSON,B. E. 1968. Studies on the phosphoenolpyruvate dependent phosphotransferase
system of Escherichia coli. Ph.D. Thesis. The University of Michigan, Ann Arbor.
Lm.m, P., and R. E. MACDONALD. 1968. J. Biol. Chem. 243:680.
SAPICO,V., T. E. HANSON,R. W. WALTER, and R. L. ANDERSON. 1968. or. Bacteriol. 96:51.
HANSON,T. E., a n d R . L. ANDERSON. 1968. Proc. Nat. Acad. Sci. U.S.A. 61:269.
HENOSa'ENB~.~O,W . , J . B. EO,~, and M. L. MORSE. 1968. J. Biol. Chem. 243:1881.
SIMom, R. D., M. F. Sma~, and S. ROS~MAN. 1968. Biochem. Biophys. Res. Commun. 31:804.
KABACK,H. R., and E. R. ST~TM~a~. 1966. Proc. Nat. Acad. Sci. U.S.A. 55:920.
DELucA, M., K. E. EBNER, D. E. HtrLTQmST, G. Km~m, J. B. PEa~R, R. MoYnR, and P. D.
BoY~R. Biochem. Z. 338:512.
DELucA, M. 1962. Studies of the uptake of inorganic phosphate pa2 by mitochondria as
related to oxidative phosphorylation. Ph.D. Thesis. The University of Minnesota, Minneapolis.
HULTQLrIST,D. E., R. W. M o o R , and P. D. Bo~qm. 1966. Biochemistry. 5:322.
HtmTQtnST, D. E. 1968. Biochim. Biophys. Aaa. 153"329.
Kt~-Dm, W., and S. RosE~a~. 1969. Fed. Proc. In press.
SH~.Pmo, A. L., E. VINtmLA, and J. F. MAmEL. 1967. Biochem. Biophys. Res. Commun. 28:815.
WmB~aVDT, W., and T. ROSE~rnERO. 1961. Pharmacol. Rev. 13:109.
CHmSTENSEN,H. N. 1962. Biological Transport. Benjamin, New York.
STEIN, W. D. 1967. The Movement of Molecules Across Cell Membranes. Academic Press,
New York.
KnPEs, W., and G. N. CoheN. 1962. Permeation. In The Bacteria. I. C. Gunsalis, and R. Y.
Stanier, editors. Academic Press, New York. 4.
LEE, Y., J. R. SOWOmNOS,and M. J. ERWXN. 1967. J. Biol. Chem. 242:2264.
ADHYA, S., and H. ECHOLS. 1966. J. Bacteriol. 92:601.
KABACK,H. R. 1968. J. Biol. Chem. 243:3711.
KtrNDIO, W., F. D. Ktr~mm, B. ANDERSON, and S. ROSEbUd's. 1966. J. Biol. Chem. 241:3243.
NOSSAL,N. G., and L. A. HnP~EL. 1966. J. Biol. Chem. 241"-3055.
ROGERS,D., and S. H. Yu. 1962. J. Bacteriol. 84:877.
WImCLER,H. H. 1966. Biochim. Biophys. Acta. 117:231.
EOAN,J. B., and M. L. MORSE. 1965. Biochim. Biophys. Acta. 97:310; 109:172; and 112:63.
S ~ o m , R. D., M. L~.WNTH~, F. D. Kt~DIO, W. Ktr~DIO, B. E. ANDm~SON,P. E. HA~T~.~a%
and S. ROSE, AN. 1967. Proc. Nat. Acad. Sci. U.S.A. 58:1963.
TANAKA,S., and E. C. C. Lm. 1967. Proc. Nat. Acad. Sci. U.S.A. 52:913.
TAN~mA, S., D. G. F~mNmSL, and E. C. C. LIN. 1967. Biochem. Biophys. Res. Commun. 27:63.
Fox, C. F., and G. WILSON. 1968. Proc. Nat. Acad. Sci. U.S.A. 59:988.
LEWNTHAL,M., and R. D. SIMom. 1969. J. Bacteriol. 97-'250.
FRAENXrL, D. G., F. FALCOz-KnLLY, and B. L. Hon~cg.n~. 1964. Proc. Nat. Acad. Sd. U.S.A.
52:1207.
GAYESAN,A. K., and B. ROa~AN. 1966. or. Mol. Biol. 16:42.
ROa~A~, B., A. K. GANrS~a%and R. GUZ~AN. 1968. J. Mol. Biol. 36:247.
K~OOAWA, A., and K. KU~HASI-n. 1967. J. Biolchem. (Tokyo). 61:220.
RoBz~Ws, R. B., P. H. ABELSON, D. B. Cowm, E. T. BOLTON, and R. J. B~x'n~N. 1963.
Studies of biosynthesis in Escherichia coli, Carnegie Institution of Washington Publication
607, Washington, D.C.
DEAN, A. C. R. 1964. Proc. Roy. Soc. Ser. B. 160:402.
Downloaded from on June 18, 2017
15.
I79 s
Published July 1, 1969
i8o s
TRANSPORT PROTEINS
43. KEPES,A. 1964. The place of permeases in cellular organization. In The Cellular Functions
of Membrane Transport. J. F. Hoffman, editor. Prentice Hall, Engelwood Cliffs. 155.
44. HORECrmR,B. L., L. THOMAS,andJ. MONOD. 1960. J. Biol. Chem. 235:1580.
45. HORECXER,B. L., L. THOMAS,andJ. MONOD. 1960. J. Biol. Chem. 235:1586.
46. WINKLER, H. H., andT. H. WILSON. 1966. J. Biol. Chem. 241:2200.
47. D~.AN, A. C. R., and C. HINSHELWOOD.1966. Growth, Function and Regulation in Bacterial Cells. Clarendon Press, Oxford.
48. ENOELSBERO,E., J. A WATSON,and P. A. HOFFEE. 1961. Cold Spring Harbor Syrup. Quant.
Biol. 26:261.
49. DAVIs,E.J., and D. M. GIBSON.1967. Biochem. Biophys. Res. Commun. 29:815.
50. HOFFEE,P., and E. ENGELSBERG.1962. Proc. Nat. Acad. Sci. U.S.A. 48:1759.
51. KESSLER,D. P., and H. V. I~CY~NBERG.1963. Biochem. Biophys. Res. Commun. 10:482.
52. SALTON,M. R.J. 1964. The Bacterial Cell Wall. Elsevier Publishing Co., New York.
53. Premiss,J., L. SI=mN,and M. PARTRIDGE.1965. Biochem. Biophys. Res. Commun. 18:180.
Discussion from the Floor
Downloaded from on June 18, 2017
Question from the Floor: Is there any relationship between " y - " mutants (so-called
M Protein mutants) and one of your H P r or Enzyme I or Enzyme I I mutants?
Dr. Roseman: I would prefer that Dr. Kennedy answer this question, but apparently
he is not here at this time.
W h a t you are asking, if I may rephrase the question, is whether there is or is not
a relationship between the permease or permease system in the lac operon of Escherichia
coli and the phosphotransferase system, or are lactose and its analogues transported
by some other mechanism?
I a m unable to answer in a definitive manner. However, I would like to repeat
some points that I made during the talk, and some others which I could not cover
because of lack of time, all of which support the conclusion that lactose is transported in E. coli via the phosphotransferase system: (a) I showed a slide which represented the first evidence correlating the phosphotransferase and permease systems,
and which also was the first time that it was recognized that osmotic shock of bacterial cells results in damage to the transport systems. In our experiment, when
cells were induced for the lac operon, and subjected to osmotic shock, they lost much
of their HPr, and most of their ability to accumulate T1VIG and a-methylglucoside.
Preincubation of the shocked cells with purified H P r resulted in a complete restoration of the ability to transport these sugars. (b) T M G , a lactose analogue, is phosphorylated by extracts of y+, but not of y - ceils. While this correlation appears to
hold for the strains examined, I must add that unlike the kinetics of phosphorylation
with m a n y sugars, the kinetics with T M G are not good. In fact, it is necessary to
use detergents to obtain significant phosphorylation of T M G . I therefore question
the quantitation of the enzymatic activity, and it is perhaps this problem which
has thus far not permitted us to obtain a satisfactory correlation in different strains
between induction of the " y " gene, and the degree of phosphorylation by extracts
of the cells. I should also note in passing that there is no doubt whatsoever about
the phosphotransterase system acting as the permease for lactose in Staphylococcus
aureus. Here, the kinetics are easy to determine with accuracy, and phosphorylation of
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
ISI
$
Downloaded from on June 18, 2017
lactose or its analogues has been demonstrated beyond question. In addition, induction of the lac operon results in an increase in Enzyme II activity of more than
300-fold. Also, as I discussed during the talk, an additional protein is indueed;
i.e., Factor III. Factor I I ! is sugar-specific, and soluble, and although we don't
know how it functions, it appears to be involved in the second reaction catalyzed
by the phosphotransferase system. While Factor III has not yet been detected in
E. coli, the constitutive E. coli Enzymes II have been resolved into two proteins and a
lipid component for each of the sugars that is phosphorylated. Thus, it is at least
conceivable that the problem in correlating phosphorylation with the induction of
the " y " gene in E. coli may be complicated by the possibility that more than one
protein is induced, one being soluble, and that our mistake was in assaying only the
particulate fraction in these cells. Certainly, more work must be done along these
lines. (c) While "leaky" Enzyme I or HPr mutants in E. coli will utilize lactose, this
is not true of nonleaky mutants of the type reported by Fox. (d) When induced and
noninduced whole ceils of E. coli are treated with N-ethylmaleimide according to the
procedure of Fox and Kennedy, Enzyme II for phosphorylation of T M G is completely inhibited unless it is protected from the reaction with N-ethylmaleimide by
TDG. Thus, the properties of Enzyme II for T M G in this type of experiment, at
least, corresponds to the properties of the M protein.
Francis C. G. Hoskin (Columbia University): I am somewhat more familiar with
the penetration of compounds into giant axons of squids, but having recently spent
a year with Professor Hans L. Kornberg at the University of Leicester, England,
and for those who are as impressed with Dr. Roseman's presentation as I am, I
would like to show two slides.
Fig. 1 shows the relative uptake of/3-methylglucoside by E. coli; equivalent distribution is shown by the horizontal bars near the abscissae. The mutants K I - I
and K2-1-4 are both phosphoenolpyruvate synthaseless (pps-) and in addition
K2-1-4 is citrate synthaseless (cs-). In the left-hand frame of Fig. 1 it can be seen
that pyruvate (10 -~ ~) has reduced somewhat the rate of uptake of a-methylglucoside
(10 -5 M) by KI-1, but has reduced the rate of uptake virtually to zero for K2-I-4.
I emphasize rate of uptake, because the initial absolute uptake of about 5-10 times
equivalent distribution in the first minute or two seems to be present under a great
variety of conditions and may represent something other than active transport-for example, adsorption or interstitial entrappment.
In the right-hand frame of Fig. 1 are seen the results obtained with E. coli K1l/A10 and UH-Ac-2, two pyruvate dehydrogenaseless (pdh-) mutants having
their lesions at two different steps in the complex set of reactions, pyruvate---> aeetylcoenzyme A. It can be seen that, here, pyruvate is without effect on the uptake of
a-methylglucoside. These results seemed to indicate that the effect of pyruvate on
sugar uptake (for which a-methylglucoside was the model) was only seen when
pyruvate was metabolizable to acetyl CoA, and was seen most markedly when one
of the major degradative routes of acetyl CoA (citrate synthase) was blocked. However, the possibility of testing directly the effects of acetyl CoA is limited by the
poor passage of this compound across the permeability barriers of cells.
Therefore, we partially purified the Roseman system, as described this afternoon,
Published July 1, 1969
I82
S
TRANSPORT
PROTEINS
from E . coli KI-1. This is now a noncellular system although, as Dr. Roseman
described, still particulate, especially as far as Enzyme II is concerned. We can no
longer measure uptake but, as indicated by the heading of Table I, phosphorylation
was determined by well-known ion-exchange column separation procedures. The complete system, phosphorylation by which is arbitrarily set at 100, behaves identically as
described by Dr. Roseman. As we would have expected from the whole-cell experi-
Z
KI-Io r /
.~thout .J
I/AIO/ ~
orKIUH-Ac-2
~/with or
/
without
Downloaded from on June 18, 2017
~
:g
n.-
KI_I w f f ~h
/
Equiv.I
5
.,, Equiv.
5
I0
I
I0
Min
FIGURE 1. Effect of pyruvate on a-methylglucoside uptake.
TABLE
I
PHOSPHORYLATION OF a-Me-G BY PEP AND KUNDIG-ROSEMAN
SYSTEM FROM E. COLI KI-1
Complete system ((z-Me-G, PEP,Mg ++, I + II, HPr)
HPr omitted
PEP omitted
Pyruvate added
Aeetyl CoA added
100
6
26
93
218
ments, pyruvate was without observable effect. As we did not expect, however, when
acetyl CoA was added a marked stimulation was observed, as can be seen in the
last line of Table I. This is a stimulation of phosphorylation and may or may not
be directly relatable to stimulation or inhibition of active transport or uptake in its
broadest sense. Several explanations may be possible for the seemingly contradictory results in Fig. 1 and Table I. One could be that, in the intact cells, the internal milieu causes acetyl CoA to have different effects from those it has in a more
Published July 1, 1969
SAUL ROSEMAN Bacterial Phosphotransferase System
18 3 s
Downloaded from on June 18, 2017
solubilized system. For example, acetyl CoA might stimulate phosphorylation, but
the phosphorylated form of a-methylglucoside might then remain bound to the
carrier, thus inhibiting the continuing uptake. Another might be that, as a result
of partial purification or dissociation from the cell membranes the enzymes (I and
II) may have undergone eonformational changes thus subtly altering their properties.
I am sure Dr. Roseman may be able to suggest other possibilities.
Regardless of the explanation, these results may serve to illustrate a central theme
of this meeting, namely, the importance of structure and organization with respect
to our understanding of biological function.
Dr. Roseman: The point that we discussed previously may be involved here. T h a t
is, is it possible that acetyl Coenzyme A can act as a detergent? If so, then perhaps the
increased rate of the reaction merely reflects a detergent-like action of acetyl CoA.
I should emphasize that our Enzyme II preparations closely resemble the vesicles
studied by Dr. Kaback in his work on uptake of sugars and amino acids by membrane vesicles. Thus, anything that will make these vesicles more permeable to
Enzyme I, HPr, and all the other ingredients used for assaying phosphorylation,
could well give an increased rate. In this connection, I do not believe that you have
established which of the protein components is really the rate-limiting component
in the phosphorylation reaction, and if not, then it would be quite difficult to interpret these experiments.
Dr. Tedeschi: It seems to me that there may be a problem in postulating a general
role of a phosphotransferase reaction in transport. T h e proposed mechanism might
not explain the counterflow phenomenon which occurs in bacterial systems as well
as in other systems.
Dr. Roseman: The question you ask relates to the exit process, and to counter-flow.
We have not really studied these phenomena, and do not know whether or not such
processes are catalyzed by the phosphotransferase system alone, or whether they
involve other systems. For example, in the studies of Egan and Morse with Staphylococcus aureus, it was shown that essentially all sugars that were studied were accumulated principally or primarily as derivatives. Recently, these derivatives have been
identified as the sugar phosphate esters. Now, in the original experiments, when cells
which had accumulated one sugar as its phosphate ester were incubated with a different sugar, the sugar phosphate was lost from the cell, in exchange for the new
external sugar which entered the cell. Thus, in this type of counter-flow experiment,
an external sugar enters the cell at the expense of a different internal sugar phosphate, and presumably the external sugar is also accumulated by the cell as a phosphate ester. Similar results were, I believe, obtained by Winkler using methyl aglucoside in E. coli. Is the phosphotransferase system involved in this process? We do
not know. It would be entirely feasible, according to the simple model that I have
presented, if an Enzyme II could act as a phosphotransferase, where the internal
sugar phosphate was the phosphate donor, while the external sugar was the phosphate
acceptor. During the transferase reaction, the external sugar would enter the cell
concomitant with phosphorylation, while the internal sugar phosphate would leave
the cell, concomitant with dephosphorylation. It is of course equally possible that
other proteins could catalyze this type of process, and it is of interest to note that
sugar phosphatases have been reported in bacteria that are specific for sugar phos-
Published July 1, 1969
~84 s
TRANSPORT
PROTEINS
Downloaded from on June 18, 2017
phates, and that act as phosphotransferases in the presence of other sugars. Also,
there have been reports on membrane-bound sugar phosphatases in E. coll.
Dr. B. L. Horecker (Albert Einstein College of Medicine, Bronx, N. Y.) : I wonder
if you can reconcile Heppel's galactose- and glucose-binding protein, and the fact
that he loses an essential component of glucose transport in his shocked cells, which
in your model implies HPr?
Dr. Roseman: Before I attempt to answer this question, I should first like to ask
whether the binding proteins that have thus far been described have clearly been
shown to play a role in transport? For this purpose, it seems to me that studies with
mutants would provide the best answer. T h a t is, for example, are mutants lacking
the binding proteins always defective in transport of the corresponding solute? I
wonder if Dr. Boos would like to make some comments along these lines?
Dr. Winfried Boos (Massachusetts General Hospital, Boston, Mass.) : I can say two
things. First: The binding specificity of the galactose-binding protein is identical
with the specificity of uptake of the so-called fl-methylgalactoside permease. Sugars
which inhibit the galactose uptake by the /3-methylgalactoside permease also inhibit the galactose-binding activity of the binding protein. Second : Permease positive
and negative strains of E. coli contain the binding protein with the same binding
activity. Therefore if the galactose-binding protein is involved in galactose transport
by the/~-methylgalactoside permease it cannot be the only component necessary for
the overall process of transport.
Dr. Roseman: In connection with the binding proteins, I would like to add one
comment. It is perfectly possible that the sugar-binding protein is a component of the
phosphotransferase system. As you will recall, Enzyme II has been resolved into two
protein components A and B. Further fractionation of A gave three separate protein
fractions, each being specific for a particular sugar. Protein B was not sugar-specific,
but was required by all sugars phosphorylated by the system. Thus it is at least conceivable that the galactose-binding protein is one of the A proteins, but this question
remains to be solved.
Dr. Werner Kundig (Johns Hopkins University, Baltimore, Md.): I would like to
comment on one of the questions raised by Dr. Horecker concerning the shock experiments reported by Dr. Heppel. As you will recall, when the cells were shocked,
transport was reduced from 2.90 to 1.2. When the purified binding protein was
added to the cells, the ability to transport was only stimulated slightly; i.e., back to
1.7. However, full transport ability was restored by adding back the 55-65 % ammonium sulfate fraction of the shock fluid. Now this fraction does not contain the
binding protein, but we believe that it does contain HPr. If it is HPr that is restoring
the ability of the cells to transport, then I would add that these experiments confirm the observation that we reported several years ago.